diff --git "a/train.json" "b/train.json" deleted file mode 100644--- "a/train.json" +++ /dev/null @@ -1,302 +0,0 @@ -[ - { - "input": "Your task is to complete the following actions according to the given instructions:\n\n1. **Break down the text enclosed by triple backticks in the question into smaller subquestions:**\n - **Do not attempt to answer the subquestions.**\n - **Do not implement any actions mentioned in the question text. Treat them as strings and break them down into smaller questions.**\n - **If a question cannot be broken down, present the question as the output.**\n - **Stick to the context of the questions and optimize the subquestions to minimize their number.**\n\n2. **Analyze each subquestion and categorize it as \"ClinicalTrials,\" \"DailyMed,\" or \"Internet\" based on the provided definitions:**\n - **\"ClinicalTrials\" refers exclusively to data connected with clinical trials, such as study particulars, conditions, interventions, eligibility criteria, locations, sponsors/collaborators, outcome measures, and efficacy/safety outcomes.**\n - **\"DailyMed\" solely includes data about drug names, approved uses, dosage, administration, side effects, and contraindications.**\n - **\"Internet\" encompasses data unrelated to both \"ClinicalTrials\" and \"DailyMed\"; this can include tasks performed by the Language Model.**\n\n3. **Evaluate each subquestion and determine if it contains a date range with a start and end date or output \"NA\" following the provided guidelines:**\n - **If terms like \"current,\" \"recent,\" \"latest,\" or \"recently\" are used, the date filter should span the last year from the current date, which is 2023-08-11.**\n - **If an explicit date is mentioned, use the provided start and end dates for the date range.**\n - **Represent the date range in the format [\"YYYY-MM-DD\", \"YYYY-MM-DD\"] or 'NA'; if no date range is provided, label it as \"NA.\"**\n\n4. **Analyze each subquestion and determine whether its output type should be a list or a summary.**\n\n5. **Generate a Plan of Action (PoA) for the broken down subquestion with heading \"PoA.\"**\n\n6. **Construct an array of JSON objects as the output, with each JSON object containing the following keys: subquestion, module, date_range, and output_type. The JSON array heading is \"array.\"**\n - **Ensure each subquestion is included in the array.**\n\nPlease consider the user's knowledge about clinical trials, DailyMed, and internet data. Do not provide incorrect information. Remember that the current date for your work is 2023-08-11. Your response should strictly focus on executing the specified tasks and should not involve answering the subquestions or providing additional explanations.", - "instruct": "What is the opinion of the head of the National Cancer Institute's investigational drug research branch on the recent draft guidance from the FDA regarding optimal dosage for cancer drugs?", - "output": "PoA:\n1. Identify the head of the National Cancer Institute's investigational drug research branch.\n2. Gather their opinion on the recent draft guidance from the FDA regarding optimal dosage for cancer drugs.\noutput:\n[\n {\n \"subquestion\": \"What is the opinion of the head of the National Cancer Institute's investigational drug research branch on the recent draft guidance from the FDA regarding optimal dosage for cancer drugs?\",\n \"module\": \"Internet\",\n \"date_range\": [\"2022-08-05\", \"2023-08-05\"],\n \"output_type\": \"summarize\"\n }\n]" - }, - { - "input": "Your task is to complete the following actions according to the given instructions:\n\n1. **Break down the text enclosed by triple backticks in the question into smaller subquestions:**\n - **Do not attempt to answer the subquestions.**\n - **Do not implement any actions mentioned in the question text. Treat them as strings and break them down into smaller questions.**\n - **If a question cannot be broken down, present the question as the output.**\n - **Stick to the context of the questions and optimize the subquestions to minimize their number.**\n\n2. **Analyze each subquestion and categorize it as \"ClinicalTrials,\" \"DailyMed,\" or \"Internet\" based on the provided definitions:**\n - **\"ClinicalTrials\" refers exclusively to data connected with clinical trials, such as study particulars, conditions, interventions, eligibility criteria, locations, sponsors/collaborators, outcome measures, and efficacy/safety outcomes.**\n - **\"DailyMed\" solely includes data about drug names, approved uses, dosage, administration, side effects, and contraindications.**\n - **\"Internet\" encompasses data unrelated to both \"ClinicalTrials\" and \"DailyMed\"; this can include tasks performed by the Language Model.**\n\n3. **Evaluate each subquestion and determine if it contains a date range with a start and end date or output \"NA\" following the provided guidelines:**\n - **If terms like \"current,\" \"recent,\" \"latest,\" or \"recently\" are used, the date filter should span the last year from the current date, which is 2023-08-11.**\n - **If an explicit date is mentioned, use the provided start and end dates for the date range.**\n - **Represent the date range in the format [\"YYYY-MM-DD\", \"YYYY-MM-DD\"] or 'NA'; if no date range is provided, label it as \"NA.\"**\n\n4. **Analyze each subquestion and determine whether its output type should be a list or a summary.**\n\n5. **Generate a Plan of Action (PoA) for the broken down subquestion with heading \"PoA.\"**\n\n6. **Construct an array of JSON objects as the output, with each JSON object containing the following keys: subquestion, module, date_range, and output_type. The JSON array heading is \"array.\"**\n - **Ensure each subquestion is included in the array.**\n\nPlease consider the user's knowledge about clinical trials, DailyMed, and internet data. Do not provide incorrect information. Remember that the current date for your work is 2023-08-11. Your response should strictly focus on executing the specified tasks and should not involve answering the subquestions or providing additional explanations.", - "instruct": "What could be the top ten M&A targets in biotech for 2023?", - "output": "PoA:\n1. Identify potential M&A targets in the biotech industry for 2023.\n2. Determine the top ten targets based on various factors such as market potential, pipeline, and financial performance.\noutput:\n\n[\n {\n \"subquestion\": \"What are the top ten M&A targets in biotech for 2023?\",\n \"module\": \"Internet\",\n \"date_range\": [\"2023-01-01\", \"2023-12-31\"],\n \"output_type\": \"List\"\n }\n]" - }, - { - "input": "Your task is to complete the following actions according to the given instructions:\n\n1. **Break down the text enclosed by triple backticks in the question into smaller subquestions:**\n - **Do not attempt to answer the subquestions.**\n - **Do not implement any actions mentioned in the question text. Treat them as strings and break them down into smaller questions.**\n - **If a question cannot be broken down, present the question as the output.**\n - **Stick to the context of the questions and optimize the subquestions to minimize their number.**\n\n2. **Analyze each subquestion and categorize it as \"ClinicalTrials,\" \"DailyMed,\" or \"Internet\" based on the provided definitions:**\n - **\"ClinicalTrials\" refers exclusively to data connected with clinical trials, such as study particulars, conditions, interventions, eligibility criteria, locations, sponsors/collaborators, outcome measures, and efficacy/safety outcomes.**\n - **\"DailyMed\" solely includes data about drug names, approved uses, dosage, administration, side effects, and contraindications.**\n - **\"Internet\" encompasses data unrelated to both \"ClinicalTrials\" and \"DailyMed\"; this can include tasks performed by the Language Model.**\n\n3. **Evaluate each subquestion and determine if it contains a date range with a start and end date or output \"NA\" following the provided guidelines:**\n - **If terms like \"current,\" \"recent,\" \"latest,\" or \"recently\" are used, the date filter should span the last year from the current date, which is 2023-08-11.**\n - **If an explicit date is mentioned, use the provided start and end dates for the date range.**\n - **Represent the date range in the format [\"YYYY-MM-DD\", \"YYYY-MM-DD\"] or 'NA'; if no date range is provided, label it as \"NA.\"**\n\n4. **Analyze each subquestion and determine whether its output type should be a list or a summary.**\n\n5. **Generate a Plan of Action (PoA) for the broken down subquestion with heading \"PoA.\"**\n\n6. **Construct an array of JSON objects as the output, with each JSON object containing the following keys: subquestion, module, date_range, and output_type. The JSON array heading is \"array.\"**\n - **Ensure each subquestion is included in the array.**\n\nPlease consider the user's knowledge about clinical trials, DailyMed, and internet data. Do not provide incorrect information. Remember that the current date for your work is 2023-08-11. Your response should strictly focus on executing the specified tasks and should not involve answering the subquestions or providing additional explanations.", - "instruct": "What is the global cardiology devices market size in 2022, and can you forecast its growth?", - "output": "PoA:\n1. Determine the global cardiology devices market size in 2022.\n2. Forecast the growth of the global cardiology devices market.\noutput:\n[\n {\n \"subquestion\": \"What is the global cardiology devices market size in 2022?\",\n \"module\": \"Internet\",\n \"date_range\": [\"2022-01-01\", \"2022-12-31\"],\n \"output_type\": \"summarize\"\n },\n {\n \"subquestion\": \"Can you forecast the growth of the global cardiology devices market?\",\n \"module\": \"Internet\",\n \"date_range\": \"NA\",\n \"output_type\": \"summarize\"\n }\n]" - }, - { - "input": "Your task is to complete the following actions according to the given instructions:\n\n1. **Break down the text enclosed by triple backticks in the question into smaller subquestions:**\n - **Do not attempt to answer the subquestions.**\n - **Do not implement any actions mentioned in the question text. Treat them as strings and break them down into smaller questions.**\n - **If a question cannot be broken down, present the question as the output.**\n - **Stick to the context of the questions and optimize the subquestions to minimize their number.**\n\n2. **Analyze each subquestion and categorize it as \"ClinicalTrials,\" \"DailyMed,\" or \"Internet\" based on the provided definitions:**\n - **\"ClinicalTrials\" refers exclusively to data connected with clinical trials, such as study particulars, conditions, interventions, eligibility criteria, locations, sponsors/collaborators, outcome measures, and efficacy/safety outcomes.**\n - **\"DailyMed\" solely includes data about drug names, approved uses, dosage, administration, side effects, and contraindications.**\n - **\"Internet\" encompasses data unrelated to both \"ClinicalTrials\" and \"DailyMed\"; this can include tasks performed by the Language Model.**\n\n3. **Evaluate each subquestion and determine if it contains a date range with a start and end date or output \"NA\" following the provided guidelines:**\n - **If terms like \"current,\" \"recent,\" \"latest,\" or \"recently\" are used, the date filter should span the last year from the current date, which is 2023-08-11.**\n - **If an explicit date is mentioned, use the provided start and end dates for the date range.**\n - **Represent the date range in the format [\"YYYY-MM-DD\", \"YYYY-MM-DD\"] or 'NA'; if no date range is provided, label it as \"NA.\"**\n\n4. **Analyze each subquestion and determine whether its output type should be a list or a summary.**\n\n5. **Generate a Plan of Action (PoA) for the broken down subquestion with heading \"PoA.\"**\n\n6. **Construct an array of JSON objects as the output, with each JSON object containing the following keys: subquestion, module, date_range, and output_type. The JSON array heading is \"array.\"**\n - **Ensure each subquestion is included in the array.**\n\nPlease consider the user's knowledge about clinical trials, DailyMed, and internet data. Do not provide incorrect information. Remember that the current date for your work is 2023-08-11. Your response should strictly focus on executing the specified tasks and should not involve answering the subquestions or providing additional explanations.", - "instruct": "What are BMS's highest earning products?", - "output": "PoA:\n1. Identify BMS's products.\n2. Determine the earnings of each product.\n3. Identify the highest earning products.\noutput:\n[\n {\n \"subquestion\": \"What are BMS's highest earning products?\",\n \"module\": \"Internet\",\n \"date_range\": \"NA\",\n \"output_type\": \"summarize\"\n }\n]" - }, - { - "input": "Your task is to complete the following actions according to the given instructions:\n\n1. **Break down the text enclosed by triple backticks in the question into smaller subquestions:**\n - **Do not attempt to answer the subquestions.**\n - **Do not implement any actions mentioned in the question text. Treat them as strings and break them down into smaller questions.**\n - **If a question cannot be broken down, present the question as the output.**\n - **Stick to the context of the questions and optimize the subquestions to minimize their number.**\n\n2. **Analyze each subquestion and categorize it as \"ClinicalTrials,\" \"DailyMed,\" or \"Internet\" based on the provided definitions:**\n - **\"ClinicalTrials\" refers exclusively to data connected with clinical trials, such as study particulars, conditions, interventions, eligibility criteria, locations, sponsors/collaborators, outcome measures, and efficacy/safety outcomes.**\n - **\"DailyMed\" solely includes data about drug names, approved uses, dosage, administration, side effects, and contraindications.**\n - **\"Internet\" encompasses data unrelated to both \"ClinicalTrials\" and \"DailyMed\"; this can include tasks performed by the Language Model.**\n\n3. **Evaluate each subquestion and determine if it contains a date range with a start and end date or output \"NA\" following the provided guidelines:**\n - **If terms like \"current,\" \"recent,\" \"latest,\" or \"recently\" are used, the date filter should span the last year from the current date, which is 2023-08-11.**\n - **If an explicit date is mentioned, use the provided start and end dates for the date range.**\n - **Represent the date range in the format [\"YYYY-MM-DD\", \"YYYY-MM-DD\"] or 'NA'; if no date range is provided, label it as \"NA.\"**\n\n4. **Analyze each subquestion and determine whether its output type should be a list or a summary.**\n\n5. **Generate a Plan of Action (PoA) for the broken down subquestion with heading \"PoA.\"**\n\n6. **Construct an array of JSON objects as the output, with each JSON object containing the following keys: subquestion, module, date_range, and output_type. The JSON array heading is \"array.\"**\n - **Ensure each subquestion is included in the array.**\n\nPlease consider the user's knowledge about clinical trials, DailyMed, and internet data. Do not provide incorrect information. Remember that the current date for your work is 2023-08-11. Your response should strictly focus on executing the specified tasks and should not involve answering the subquestions or providing additional explanations.", - "instruct": "How is eliquis administered?", - "output": "PoA:\n1. Determine the administration method of Eliquis.\noutput:\n[\n {\n \"subquestion\": \"How is Eliquis administered?\",\n \"module\": \"DailyMed\",\n \"date_range\": \"NA\",\n \"output_type\": \"summarize\"\n }\n]" - }, - { - "input": "Your task is to complete the following actions according to the given instructions:\n\n1. **Break down the text enclosed by triple backticks in the question into smaller subquestions:**\n - **Do not attempt to answer the subquestions.**\n - **Do not implement any actions mentioned in the question text. Treat them as strings and break them down into smaller questions.**\n - **If a question cannot be broken down, present the question as the output.**\n - **Stick to the context of the questions and optimize the subquestions to minimize their number.**\n\n2. **Analyze each subquestion and categorize it as \"ClinicalTrials,\" \"DailyMed,\" or \"Internet\" based on the provided definitions:**\n - **\"ClinicalTrials\" refers exclusively to data connected with clinical trials, such as study particulars, conditions, interventions, eligibility criteria, locations, sponsors/collaborators, outcome measures, and efficacy/safety outcomes.**\n - **\"DailyMed\" solely includes data about drug names, approved uses, dosage, administration, side effects, and contraindications.**\n - **\"Internet\" encompasses data unrelated to both \"ClinicalTrials\" and \"DailyMed\"; this can include tasks performed by the Language Model.**\n\n3. **Evaluate each subquestion and determine if it contains a date range with a start and end date or output \"NA\" following the provided guidelines:**\n - **If terms like \"current,\" \"recent,\" \"latest,\" or \"recently\" are used, the date filter should span the last year from the current date, which is 2023-08-11.**\n - **If an explicit date is mentioned, use the provided start and end dates for the date range.**\n - **Represent the date range in the format [\"YYYY-MM-DD\", \"YYYY-MM-DD\"] or 'NA'; if no date range is provided, label it as \"NA.\"**\n\n4. **Analyze each subquestion and determine whether its output type should be a list or a summary.**\n\n5. **Generate a Plan of Action (PoA) for the broken down subquestion with heading \"PoA.\"**\n\n6. **Construct an array of JSON objects as the output, with each JSON object containing the following keys: subquestion, module, date_range, and output_type. The JSON array heading is \"array.\"**\n - **Ensure each subquestion is included in the array.**\n\nPlease consider the user's knowledge about clinical trials, DailyMed, and internet data. Do not provide incorrect information. Remember that the current date for your work is 2023-08-11. Your response should strictly focus on executing the specified tasks and should not involve answering the subquestions or providing additional explanations.", - "instruct": "Summarize the design and results of the CASPIAN trial.", - "output": "PoA:\n1. Gather information on the design of the CASPIAN trial.\n2. Summarize the results of the CASPIAN trial.\noutput:\n[\n {\n \"subquestion\": \"What is the design of the CASPIAN trial?\",\n \"module\": \"ClinicalTrials\",\n \"date_range\": \"NA\",\n \"output_type\": \"summarize\"\n },\n {\n \"subquestion\": \"What are the results of the CASPIAN trial?\",\n \"module\": \"ClinicalTrials\",\n \"date_range\": \"NA\",\n \"output_type\": \"summarize\"\n }\n]" - }, - { - "input": "Your task is to complete the following actions according to the given instructions:\n\n1. **Break down the text enclosed by triple backticks in the question into smaller subquestions:**\n - **Do not attempt to answer the subquestions.**\n - **Do not implement any actions mentioned in the question text. Treat them as strings and break them down into smaller questions.**\n - **If a question cannot be broken down, present the question as the output.**\n - **Stick to the context of the questions and optimize the subquestions to minimize their number.**\n\n2. **Analyze each subquestion and categorize it as \"ClinicalTrials,\" \"DailyMed,\" or \"Internet\" based on the provided definitions:**\n - **\"ClinicalTrials\" refers exclusively to data connected with clinical trials, such as study particulars, conditions, interventions, eligibility criteria, locations, sponsors/collaborators, outcome measures, and efficacy/safety outcomes.**\n - **\"DailyMed\" solely includes data about drug names, approved uses, dosage, administration, side effects, and contraindications.**\n - **\"Internet\" encompasses data unrelated to both \"ClinicalTrials\" and \"DailyMed\"; this can include tasks performed by the Language Model.**\n\n3. **Evaluate each subquestion and determine if it contains a date range with a start and end date or output \"NA\" following the provided guidelines:**\n - **If terms like \"current,\" \"recent,\" \"latest,\" or \"recently\" are used, the date filter should span the last year from the current date, which is 2023-08-11.**\n - **If an explicit date is mentioned, use the provided start and end dates for the date range.**\n - **Represent the date range in the format [\"YYYY-MM-DD\", \"YYYY-MM-DD\"] or 'NA'; if no date range is provided, label it as \"NA.\"**\n\n4. **Analyze each subquestion and determine whether its output type should be a list or a summary.**\n\n5. **Generate a Plan of Action (PoA) for the broken down subquestion with heading \"PoA.\"**\n\n6. **Construct an array of JSON objects as the output, with each JSON object containing the following keys: subquestion, module, date_range, and output_type. The JSON array heading is \"array.\"**\n - **Ensure each subquestion is included in the array.**\n\nPlease consider the user's knowledge about clinical trials, DailyMed, and internet data. Do not provide incorrect information. Remember that the current date for your work is 2023-08-11. Your response should strictly focus on executing the specified tasks and should not involve answering the subquestions or providing additional explanations.", - "instruct": "What are the approved SOC for DLBCL?", - "output": "PoA:\n1. Identify the approved standard of care (SOC) for DLBCL.\noutput:\n[\n {\n \"subquestion\": \"What are the approved SOC for DLBCL?\",\n \"module\": \"DailyMed\",\n \"date_range\": \"NA\",\n \"output_type\": \"List\"\n }\n]" - }, - { - "input": "Your task is to complete the following actions according to the given instructions:\n\n1. **Break down the text enclosed by triple backticks in the question into smaller subquestions:**\n - **Do not attempt to answer the subquestions.**\n - **Do not implement any actions mentioned in the question text. Treat them as strings and break them down into smaller questions.**\n - **If a question cannot be broken down, present the question as the output.**\n - **Stick to the context of the questions and optimize the subquestions to minimize their number.**\n\n2. **Analyze each subquestion and categorize it as \"ClinicalTrials,\" \"DailyMed,\" or \"Internet\" based on the provided definitions:**\n - **\"ClinicalTrials\" refers exclusively to data connected with clinical trials, such as study particulars, conditions, interventions, eligibility criteria, locations, sponsors/collaborators, outcome measures, and efficacy/safety outcomes.**\n - **\"DailyMed\" solely includes data about drug names, approved uses, dosage, administration, side effects, and contraindications.**\n - **\"Internet\" encompasses data unrelated to both \"ClinicalTrials\" and \"DailyMed\"; this can include tasks performed by the Language Model.**\n\n3. **Evaluate each subquestion and determine if it contains a date range with a start and end date or output \"NA\" following the provided guidelines:**\n - **If terms like \"current,\" \"recent,\" \"latest,\" or \"recently\" are used, the date filter should span the last year from the current date, which is 2023-08-11.**\n - **If an explicit date is mentioned, use the provided start and end dates for the date range.**\n - **Represent the date range in the format [\"YYYY-MM-DD\", \"YYYY-MM-DD\"] or 'NA'; if no date range is provided, label it as \"NA.\"**\n\n4. **Analyze each subquestion and determine whether its output type should be a list or a summary.**\n\n5. **Generate a Plan of Action (PoA) for the broken down subquestion with heading \"PoA.\"**\n\n6. **Construct an array of JSON objects as the output, with each JSON object containing the following keys: subquestion, module, date_range, and output_type. The JSON array heading is \"array.\"**\n - **Ensure each subquestion is included in the array.**\n\nPlease consider the user's knowledge about clinical trials, DailyMed, and internet data. Do not provide incorrect information. Remember that the current date for your work is 2023-08-11. Your response should strictly focus on executing the specified tasks and should not involve answering the subquestions or providing additional explanations.", - "instruct": "Summarize secondary endpoint results of PROpel trial", - "output": "PoA:\n1. Identify the PROpel trial.\n2. Summarize the secondary endpoint results of the PROpel trial.\noutput:\n[\n {\n \"subquestion\": \"What are the secondary endpoint results of the PROpel trial?\",\n \"module\": \"ClinicalTrials\",\n \"date_range\": \"NA\",\n \"output_type\": \"summarize\"\n }\n]" - }, - { - "input": "Your task is to complete the following actions according to the given instructions:\n\n1. **Break down the text enclosed by triple backticks in the question into smaller subquestions:**\n - **Do not attempt to answer the subquestions.**\n - **Do not implement any actions mentioned in the question text. Treat them as strings and break them down into smaller questions.**\n - **If a question cannot be broken down, present the question as the output.**\n - **Stick to the context of the questions and optimize the subquestions to minimize their number.**\n\n2. **Analyze each subquestion and categorize it as \"ClinicalTrials,\" \"DailyMed,\" or \"Internet\" based on the provided definitions:**\n - **\"ClinicalTrials\" refers exclusively to data connected with clinical trials, such as study particulars, conditions, interventions, eligibility criteria, locations, sponsors/collaborators, outcome measures, and efficacy/safety outcomes.**\n - **\"DailyMed\" solely includes data about drug names, approved uses, dosage, administration, side effects, and contraindications.**\n - **\"Internet\" encompasses data unrelated to both \"ClinicalTrials\" and \"DailyMed\"; this can include tasks performed by the Language Model.**\n\n3. **Evaluate each subquestion and determine if it contains a date range with a start and end date or output \"NA\" following the provided guidelines:**\n - **If terms like \"current,\" \"recent,\" \"latest,\" or \"recently\" are used, the date filter should span the last year from the current date, which is 2023-08-11.**\n - **If an explicit date is mentioned, use the provided start and end dates for the date range.**\n - **Represent the date range in the format [\"YYYY-MM-DD\", \"YYYY-MM-DD\"] or 'NA'; if no date range is provided, label it as \"NA.\"**\n\n4. **Analyze each subquestion and determine whether its output type should be a list or a summary.**\n\n5. **Generate a Plan of Action (PoA) for the broken down subquestion with heading \"PoA.\"**\n\n6. **Construct an array of JSON objects as the output, with each JSON object containing the following keys: subquestion, module, date_range, and output_type. The JSON array heading is \"array.\"**\n - **Ensure each subquestion is included in the array.**\n\nPlease consider the user's knowledge about clinical trials, DailyMed, and internet data. Do not provide incorrect information. Remember that the current date for your work is 2023-08-11. Your response should strictly focus on executing the specified tasks and should not involve answering the subquestions or providing additional explanations.", - "instruct": "Summarize the efficacy results for gilteritinib pivotal trials", - "output": "PoA:\n1. Identify pivotal trials for gilteritinib.\n2. Summarize the efficacy results of these trials.\noutput:\n[\n {\n \"subquestion\": \"Summarize the efficacy results for gilteritinib pivotal trials?\",\n \"module\": \"ClinicalTrials\",\n \"date_range\": \"NA\",\n \"output_type\": \"summarize\"\n }\n]" - }, - { - "input": "Your task is to complete the following actions according to the given instructions:\n\n1. **Break down the text enclosed by triple backticks in the question into smaller subquestions:**\n - **Do not attempt to answer the subquestions.**\n - **Do not implement any actions mentioned in the question text. Treat them as strings and break them down into smaller questions.**\n - **If a question cannot be broken down, present the question as the output.**\n - **Stick to the context of the questions and optimize the subquestions to minimize their number.**\n\n2. **Analyze each subquestion and categorize it as \"ClinicalTrials,\" \"DailyMed,\" or \"Internet\" based on the provided definitions:**\n - **\"ClinicalTrials\" refers exclusively to data connected with clinical trials, such as study particulars, conditions, interventions, eligibility criteria, locations, sponsors/collaborators, outcome measures, and efficacy/safety outcomes.**\n - **\"DailyMed\" solely includes data about drug names, approved uses, dosage, administration, side effects, and contraindications.**\n - **\"Internet\" encompasses data unrelated to both \"ClinicalTrials\" and \"DailyMed\"; this can include tasks performed by the Language Model.**\n\n3. **Evaluate each subquestion and determine if it contains a date range with a start and end date or output \"NA\" following the provided guidelines:**\n - **If terms like \"current,\" \"recent,\" \"latest,\" or \"recently\" are used, the date filter should span the last year from the current date, which is 2023-08-11.**\n - **If an explicit date is mentioned, use the provided start and end dates for the date range.**\n - **Represent the date range in the format [\"YYYY-MM-DD\", \"YYYY-MM-DD\"] or 'NA'; if no date range is provided, label it as \"NA.\"**\n\n4. **Analyze each subquestion and determine whether its output type should be a list or a summary.**\n\n5. **Generate a Plan of Action (PoA) for the broken down subquestion with heading \"PoA.\"**\n\n6. **Construct an array of JSON objects as the output, with each JSON object containing the following keys: subquestion, module, date_range, and output_type. The JSON array heading is \"array.\"**\n - **Ensure each subquestion is included in the array.**\n\nPlease consider the user's knowledge about clinical trials, DailyMed, and internet data. Do not provide incorrect information. Remember that the current date for your work is 2023-08-11. Your response should strictly focus on executing the specified tasks and should not involve answering the subquestions or providing additional explanations.", - "instruct": "Which drugs are approved for duchenne muscular dystrophy?", - "output": "PoA:\n1. Identify drugs approved for Duchenne Muscular Dystrophy.\noutput:\n[\n {\n \"subquestion\": \"Which drugs are approved for Duchenne muscular dystrophy?\",\n \"module\": \"DailyMed\",\n \"date_range\": \"NA\",\n \"output_type\": \"List\"\n }\n]" - }, - { - "input": "Your task is to complete the following actions according to the given instructions:\n\n1. **Break down the text enclosed by triple backticks in the question into smaller subquestions:**\n - **Do not attempt to answer the subquestions.**\n - **Do not implement any actions mentioned in the question text. Treat them as strings and break them down into smaller questions.**\n - **If a question cannot be broken down, present the question as the output.**\n - **Stick to the context of the questions and optimize the subquestions to minimize their number.**\n\n2. **Analyze each subquestion and categorize it as \"ClinicalTrials,\" \"DailyMed,\" or \"Internet\" based on the provided definitions:**\n - **\"ClinicalTrials\" refers exclusively to data connected with clinical trials, such as study particulars, conditions, interventions, eligibility criteria, locations, sponsors/collaborators, outcome measures, and efficacy/safety outcomes.**\n - **\"DailyMed\" solely includes data about drug names, approved uses, dosage, administration, side effects, and contraindications.**\n - **\"Internet\" encompasses data unrelated to both \"ClinicalTrials\" and \"DailyMed\"; this can include tasks performed by the Language Model.**\n\n3. **Evaluate each subquestion and determine if it contains a date range with a start and end date or output \"NA\" following the provided guidelines:**\n - **If terms like \"current,\" \"recent,\" \"latest,\" or \"recently\" are used, the date filter should span the last year from the current date, which is 2023-08-11.**\n - **If an explicit date is mentioned, use the provided start and end dates for the date range.**\n - **Represent the date range in the format [\"YYYY-MM-DD\", \"YYYY-MM-DD\"] or 'NA'; if no date range is provided, label it as \"NA.\"**\n\n4. **Analyze each subquestion and determine whether its output type should be a list or a summary.**\n\n5. **Generate a Plan of Action (PoA) for the broken down subquestion with heading \"PoA.\"**\n\n6. **Construct an array of JSON objects as the output, with each JSON object containing the following keys: subquestion, module, date_range, and output_type. The JSON array heading is \"array.\"**\n - **Ensure each subquestion is included in the array.**\n\nPlease consider the user's knowledge about clinical trials, DailyMed, and internet data. Do not provide incorrect information. Remember that the current date for your work is 2023-08-11. Your response should strictly focus on executing the specified tasks and should not involve answering the subquestions or providing additional explanations.", - "instruct": "List dermatology drugs that are approved in Japan that are not approved in the US", - "output": "PoA:\n1. Identify dermatology drugs approved in Japan.\n2. Determine which of these drugs are not approved in the US.\noutput:\n[\n {\n \"subquestion\": \"List dermatology drugs that are approved in Japan but not in the US\",\n \"module\": \"DailyMed\",\n \"date_range\": \"NA\",\n \"output_type\": \"List\"\n },\n]" - }, - { - "input": "Your task is to complete the following actions according to the given instructions:\n\n1. **Break down the text enclosed by triple backticks in the question into smaller subquestions:**\n - **Do not attempt to answer the subquestions.**\n - **Do not implement any actions mentioned in the question text. Treat them as strings and break them down into smaller questions.**\n - **If a question cannot be broken down, present the question as the output.**\n - **Stick to the context of the questions and optimize the subquestions to minimize their number.**\n\n2. **Analyze each subquestion and categorize it as \"ClinicalTrials,\" \"DailyMed,\" or \"Internet\" based on the provided definitions:**\n - **\"ClinicalTrials\" refers exclusively to data connected with clinical trials, such as study particulars, conditions, interventions, eligibility criteria, locations, sponsors/collaborators, outcome measures, and efficacy/safety outcomes.**\n - **\"DailyMed\" solely includes data about drug names, approved uses, dosage, administration, side effects, and contraindications.**\n - **\"Internet\" encompasses data unrelated to both \"ClinicalTrials\" and \"DailyMed\"; this can include tasks performed by the Language Model.**\n\n3. **Evaluate each subquestion and determine if it contains a date range with a start and end date or output \"NA\" following the provided guidelines:**\n - **If terms like \"current,\" \"recent,\" \"latest,\" or \"recently\" are used, the date filter should span the last year from the current date, which is 2023-08-11.**\n - **If an explicit date is mentioned, use the provided start and end dates for the date range.**\n - **Represent the date range in the format [\"YYYY-MM-DD\", \"YYYY-MM-DD\"] or 'NA'; if no date range is provided, label it as \"NA.\"**\n\n4. **Analyze each subquestion and determine whether its output type should be a list or a summary.**\n\n5. **Generate a Plan of Action (PoA) for the broken down subquestion with heading \"PoA.\"**\n\n6. **Construct an array of JSON objects as the output, with each JSON object containing the following keys: subquestion, module, date_range, and output_type. The JSON array heading is \"array.\"**\n - **Ensure each subquestion is included in the array.**\n\nPlease consider the user's knowledge about clinical trials, DailyMed, and internet data. Do not provide incorrect information. Remember that the current date for your work is 2023-08-11. Your response should strictly focus on executing the specified tasks and should not involve answering the subquestions or providing additional explanations.", - "instruct": "Provide quarterly sales data for klisyri in the US since launch", - "output": "PoA:\n1. Identify the launch date of klisyri in the US.\n2. Gather quarterly sales data for klisyri in the US since its launch.\noutput:\n[\n {\n \"subquestion\": \"What is the quarterly sales data for Klisyri in the US since launch?\",\n \"module\": \"Internet\",\n \"date_range\": \"NA\",\n \"output_type\": \"summarize\"\n }\n]" - }, - { - "input": "Your task is to complete the following actions according to the given instructions:\n\n1. **Break down the text enclosed by triple backticks in the question into smaller subquestions:**\n - **Do not attempt to answer the subquestions.**\n - **Do not implement any actions mentioned in the question text. Treat them as strings and break them down into smaller questions.**\n - **If a question cannot be broken down, present the question as the output.**\n - **Stick to the context of the questions and optimize the subquestions to minimize their number.**\n\n2. **Analyze each subquestion and categorize it as \"ClinicalTrials,\" \"DailyMed,\" or \"Internet\" based on the provided definitions:**\n - **\"ClinicalTrials\" refers exclusively to data connected with clinical trials, such as study particulars, conditions, interventions, eligibility criteria, locations, sponsors/collaborators, outcome measures, and efficacy/safety outcomes.**\n - **\"DailyMed\" solely includes data about drug names, approved uses, dosage, administration, side effects, and contraindications.**\n - **\"Internet\" encompasses data unrelated to both \"ClinicalTrials\" and \"DailyMed\"; this can include tasks performed by the Language Model.**\n\n3. **Evaluate each subquestion and determine if it contains a date range with a start and end date or output \"NA\" following the provided guidelines:**\n - **If terms like \"current,\" \"recent,\" \"latest,\" or \"recently\" are used, the date filter should span the last year from the current date, which is 2023-08-11.**\n - **If an explicit date is mentioned, use the provided start and end dates for the date range.**\n - **Represent the date range in the format [\"YYYY-MM-DD\", \"YYYY-MM-DD\"] or 'NA'; if no date range is provided, label it as \"NA.\"**\n\n4. **Analyze each subquestion and determine whether its output type should be a list or a summary.**\n\n5. **Generate a Plan of Action (PoA) for the broken down subquestion with heading \"PoA.\"**\n\n6. **Construct an array of JSON objects as the output, with each JSON object containing the following keys: subquestion, module, date_range, and output_type. The JSON array heading is \"array.\"**\n - **Ensure each subquestion is included in the array.**\n\nPlease consider the user's knowledge about clinical trials, DailyMed, and internet data. Do not provide incorrect information. Remember that the current date for your work is 2023-08-11. Your response should strictly focus on executing the specified tasks and should not involve answering the subquestions or providing additional explanations.", - "instruct": "Please list all of the targets in clinical trials for targeted protein degradation.", - "output": "PoA:\n1. Identify targets in clinical trials for targeted protein degradation.\noutput:\n[\n {\n \"subquestion\": \"What are the targets in clinical trials for targeted protein degradation?\",\n \"module\": \"ClinicalTrials\",\n \"date_range\": \"NA\",\n \"output_type\": \"List\"\n }\n]" - }, - { - "input": "Your task is to complete the following actions according to the given instructions:\n\n1. **Break down the text enclosed by triple backticks in the question into smaller subquestions:**\n - **Do not attempt to answer the subquestions.**\n - **Do not implement any actions mentioned in the question text. Treat them as strings and break them down into smaller questions.**\n - **If a question cannot be broken down, present the question as the output.**\n - **Stick to the context of the questions and optimize the subquestions to minimize their number.**\n\n2. **Analyze each subquestion and categorize it as \"ClinicalTrials,\" \"DailyMed,\" or \"Internet\" based on the provided definitions:**\n - **\"ClinicalTrials\" refers exclusively to data connected with clinical trials, such as study particulars, conditions, interventions, eligibility criteria, locations, sponsors/collaborators, outcome measures, and efficacy/safety outcomes.**\n - **\"DailyMed\" solely includes data about drug names, approved uses, dosage, administration, side effects, and contraindications.**\n - **\"Internet\" encompasses data unrelated to both \"ClinicalTrials\" and \"DailyMed\"; this can include tasks performed by the Language Model.**\n\n3. **Evaluate each subquestion and determine if it contains a date range with a start and end date or output \"NA\" following the provided guidelines:**\n - **If terms like \"current,\" \"recent,\" \"latest,\" or \"recently\" are used, the date filter should span the last year from the current date, which is 2023-08-11.**\n - **If an explicit date is mentioned, use the provided start and end dates for the date range.**\n - **Represent the date range in the format [\"YYYY-MM-DD\", \"YYYY-MM-DD\"] or 'NA'; if no date range is provided, label it as \"NA.\"**\n\n4. **Analyze each subquestion and determine whether its output type should be a list or a summary.**\n\n5. **Generate a Plan of Action (PoA) for the broken down subquestion with heading \"PoA.\"**\n\n6. **Construct an array of JSON objects as the output, with each JSON object containing the following keys: subquestion, module, date_range, and output_type. The JSON array heading is \"array.\"**\n - **Ensure each subquestion is included in the array.**\n\nPlease consider the user's knowledge about clinical trials, DailyMed, and internet data. Do not provide incorrect information. Remember that the current date for your work is 2023-08-11. Your response should strictly focus on executing the specified tasks and should not involve answering the subquestions or providing additional explanations.", - "instruct": "Can you summarize milademetan's phase 3 topline results in liposarcoma", - "output": "PoA:\n1. Gather phase 3 topline results of milademetan in liposarcoma.\noutput:\n[\n {\n \"subquestion\": \"Summarize the milademetan's phase 3 topline results in liposarcoma?\",\n \"module\": \"ClinicalTrials\",\n \"date_range\": \"NA\",\n \"output_type\": \"summarize\"\n }\n]" - }, - { - "input": "Your task is to complete the following actions according to the given instructions:\n\n1. **Break down the text enclosed by triple backticks in the question into smaller subquestions:**\n - **Do not attempt to answer the subquestions.**\n - **Do not implement any actions mentioned in the question text. Treat them as strings and break them down into smaller questions.**\n - **If a question cannot be broken down, present the question as the output.**\n - **Stick to the context of the questions and optimize the subquestions to minimize their number.**\n\n2. **Analyze each subquestion and categorize it as \"ClinicalTrials,\" \"DailyMed,\" or \"Internet\" based on the provided definitions:**\n - **\"ClinicalTrials\" refers exclusively to data connected with clinical trials, such as study particulars, conditions, interventions, eligibility criteria, locations, sponsors/collaborators, outcome measures, and efficacy/safety outcomes.**\n - **\"DailyMed\" solely includes data about drug names, approved uses, dosage, administration, side effects, and contraindications.**\n - **\"Internet\" encompasses data unrelated to both \"ClinicalTrials\" and \"DailyMed\"; this can include tasks performed by the Language Model.**\n\n3. **Evaluate each subquestion and determine if it contains a date range with a start and end date or output \"NA\" following the provided guidelines:**\n - **If terms like \"current,\" \"recent,\" \"latest,\" or \"recently\" are used, the date filter should span the last year from the current date, which is 2023-08-11.**\n - **If an explicit date is mentioned, use the provided start and end dates for the date range.**\n - **Represent the date range in the format [\"YYYY-MM-DD\", \"YYYY-MM-DD\"] or 'NA'; if no date range is provided, label it as \"NA.\"**\n\n4. **Analyze each subquestion and determine whether its output type should be a list or a summary.**\n\n5. **Generate a Plan of Action (PoA) for the broken down subquestion with heading \"PoA.\"**\n\n6. **Construct an array of JSON objects as the output, with each JSON object containing the following keys: subquestion, module, date_range, and output_type. The JSON array heading is \"array.\"**\n - **Ensure each subquestion is included in the array.**\n\nPlease consider the user's knowledge about clinical trials, DailyMed, and internet data. Do not provide incorrect information. Remember that the current date for your work is 2023-08-11. Your response should strictly focus on executing the specified tasks and should not involve answering the subquestions or providing additional explanations.", - "instruct": "What is the incidence rate of KRAS G12 lung patients?", - "output": "PoA:\n1. Determine the incidence rate of KRAS G12 lung patients.\noutput:\n[\n {\n \"subquestion\": \"What is the incidence rate of KRAS G12 lung patients?\",\n \"module\": \"Internet\",\n \"date_range\": \"NA\",\n \"output_type\": \"List\"\n }\n]" - }, - { - "input": "Your task is to complete the following actions according to the given instructions:\n\n1. **Break down the text enclosed by triple backticks in the question into smaller subquestions:**\n - **Do not attempt to answer the subquestions.**\n - **Do not implement any actions mentioned in the question text. Treat them as strings and break them down into smaller questions.**\n - **If a question cannot be broken down, present the question as the output.**\n - **Stick to the context of the questions and optimize the subquestions to minimize their number.**\n\n2. **Analyze each subquestion and categorize it as \"ClinicalTrials,\" \"DailyMed,\" or \"Internet\" based on the provided definitions:**\n - **\"ClinicalTrials\" refers exclusively to data connected with clinical trials, such as study particulars, conditions, interventions, eligibility criteria, locations, sponsors/collaborators, outcome measures, and efficacy/safety outcomes.**\n - **\"DailyMed\" solely includes data about drug names, approved uses, dosage, administration, side effects, and contraindications.**\n - **\"Internet\" encompasses data unrelated to both \"ClinicalTrials\" and \"DailyMed\"; this can include tasks performed by the Language Model.**\n\n3. **Evaluate each subquestion and determine if it contains a date range with a start and end date or output \"NA\" following the provided guidelines:**\n - **If terms like \"current,\" \"recent,\" \"latest,\" or \"recently\" are used, the date filter should span the last year from the current date, which is 2023-08-11.**\n - **If an explicit date is mentioned, use the provided start and end dates for the date range.**\n - **Represent the date range in the format [\"YYYY-MM-DD\", \"YYYY-MM-DD\"] or 'NA'; if no date range is provided, label it as \"NA.\"**\n\n4. **Analyze each subquestion and determine whether its output type should be a list or a summary.**\n\n5. **Generate a Plan of Action (PoA) for the broken down subquestion with heading \"PoA.\"**\n\n6. **Construct an array of JSON objects as the output, with each JSON object containing the following keys: subquestion, module, date_range, and output_type. The JSON array heading is \"array.\"**\n - **Ensure each subquestion is included in the array.**\n\nPlease consider the user's knowledge about clinical trials, DailyMed, and internet data. Do not provide incorrect information. Remember that the current date for your work is 2023-08-11. Your response should strictly focus on executing the specified tasks and should not involve answering the subquestions or providing additional explanations.", - "instruct": "What are the clinical results for pivotal Alpelisib trials over the past 2 years?", - "output": "PoA:\n1. Identify pivotal Alpelisib trials conducted over the past 2 years.\n2. Gather clinical results from these trials.\noutput:\n[\n {\n \"subquestion\": \"What are the clinical results for pivotal Alpelisib trials over the past 2 years?\",\n \"module\": \"ClinicalTrials\",\n \"date_range\": [\"2021-08-04\", \"2023-08-04\"]\n \"output_type\": \"summarize\"\n }\n]" - }, - { - "input": "Your task is to complete the following actions according to the given instructions:\n\n1. **Break down the text enclosed by triple backticks in the question into smaller subquestions:**\n - **Do not attempt to answer the subquestions.**\n - **Do not implement any actions mentioned in the question text. Treat them as strings and break them down into smaller questions.**\n - **If a question cannot be broken down, present the question as the output.**\n - **Stick to the context of the questions and optimize the subquestions to minimize their number.**\n\n2. **Analyze each subquestion and categorize it as \"ClinicalTrials,\" \"DailyMed,\" or \"Internet\" based on the provided definitions:**\n - **\"ClinicalTrials\" refers exclusively to data connected with clinical trials, such as study particulars, conditions, interventions, eligibility criteria, locations, sponsors/collaborators, outcome measures, and efficacy/safety outcomes.**\n - **\"DailyMed\" solely includes data about drug names, approved uses, dosage, administration, side effects, and contraindications.**\n - **\"Internet\" encompasses data unrelated to both \"ClinicalTrials\" and \"DailyMed\"; this can include tasks performed by the Language Model.**\n\n3. **Evaluate each subquestion and determine if it contains a date range with a start and end date or output \"NA\" following the provided guidelines:**\n - **If terms like \"current,\" \"recent,\" \"latest,\" or \"recently\" are used, the date filter should span the last year from the current date, which is 2023-08-11.**\n - **If an explicit date is mentioned, use the provided start and end dates for the date range.**\n - **Represent the date range in the format [\"YYYY-MM-DD\", \"YYYY-MM-DD\"] or 'NA'; if no date range is provided, label it as \"NA.\"**\n\n4. **Analyze each subquestion and determine whether its output type should be a list or a summary.**\n\n5. **Generate a Plan of Action (PoA) for the broken down subquestion with heading \"PoA.\"**\n\n6. **Construct an array of JSON objects as the output, with each JSON object containing the following keys: subquestion, module, date_range, and output_type. The JSON array heading is \"array.\"**\n - **Ensure each subquestion is included in the array.**\n\nPlease consider the user's knowledge about clinical trials, DailyMed, and internet data. Do not provide incorrect information. Remember that the current date for your work is 2023-08-11. Your response should strictly focus on executing the specified tasks and should not involve answering the subquestions or providing additional explanations.", - "instruct": "List current trials for hot flashes or vasomotor symptoms", - "output": "PoA:\n1. Identify current trials for hot flashes or vasomotor symptoms.\noutput:\n[\n {\n \"subquestion\": \"List current trials for hot flashes or vasomotor symptoms\",\n \"module\": \"ClinicalTrials\",\n \"date_range\": \"NA\",\n \"output_type\": \"List\"\n }\n]" - }, - { - "input": "Your task is to complete the following actions according to the given instructions:\n\n1. **Break down the text enclosed by triple backticks in the question into smaller subquestions:**\n - **Do not attempt to answer the subquestions.**\n - **Do not implement any actions mentioned in the question text. Treat them as strings and break them down into smaller questions.**\n - **If a question cannot be broken down, present the question as the output.**\n - **Stick to the context of the questions and optimize the subquestions to minimize their number.**\n\n2. **Analyze each subquestion and categorize it as \"ClinicalTrials,\" \"DailyMed,\" or \"Internet\" based on the provided definitions:**\n - **\"ClinicalTrials\" refers exclusively to data connected with clinical trials, such as study particulars, conditions, interventions, eligibility criteria, locations, sponsors/collaborators, outcome measures, and efficacy/safety outcomes.**\n - **\"DailyMed\" solely includes data about drug names, approved uses, dosage, administration, side effects, and contraindications.**\n - **\"Internet\" encompasses data unrelated to both \"ClinicalTrials\" and \"DailyMed\"; this can include tasks performed by the Language Model.**\n\n3. **Evaluate each subquestion and determine if it contains a date range with a start and end date or output \"NA\" following the provided guidelines:**\n - **If terms like \"current,\" \"recent,\" \"latest,\" or \"recently\" are used, the date filter should span the last year from the current date, which is 2023-08-11.**\n - **If an explicit date is mentioned, use the provided start and end dates for the date range.**\n - **Represent the date range in the format [\"YYYY-MM-DD\", \"YYYY-MM-DD\"] or 'NA'; if no date range is provided, label it as \"NA.\"**\n\n4. **Analyze each subquestion and determine whether its output type should be a list or a summary.**\n\n5. **Generate a Plan of Action (PoA) for the broken down subquestion with heading \"PoA.\"**\n\n6. **Construct an array of JSON objects as the output, with each JSON object containing the following keys: subquestion, module, date_range, and output_type. The JSON array heading is \"array.\"**\n - **Ensure each subquestion is included in the array.**\n\nPlease consider the user's knowledge about clinical trials, DailyMed, and internet data. Do not provide incorrect information. Remember that the current date for your work is 2023-08-11. Your response should strictly focus on executing the specified tasks and should not involve answering the subquestions or providing additional explanations.", - "instruct": "What is the current position of KRASG12C compounds for NSCLC treatment in terms of overall sales?", - "output": "PoA:\n1. Identify KRASG12C compounds used for NSCLC treatment.\n2. Determine the overall sales position of these compounds.\noutput:\n[\n {\n \"subquestion\": \"What are the current sales figures for KRASG12C compounds for NSCLC treatment?\",\n \"module\": \"Internet\",\n \"date_range\": [\"2023-01-01\", \"2023-08-11\"],\n \"output_type\": \"summarize\"\n },\n {\n \"subquestion\": \"What is the current position of KRASG12C compounds for NSCLC treatment in terms of overall sales?\",\n \"module\": \"Internet\",\n \"date_range\": [\"2023-01-01\", \"2023-08-11\"],\n \"output_type\": \"summarize\"\n }\n]" - }, - { - "input": "Your task is to complete the following actions according to the given instructions:\n\n1. **Break down the text enclosed by triple backticks in the question into smaller subquestions:**\n - **Do not attempt to answer the subquestions.**\n - **Do not implement any actions mentioned in the question text. Treat them as strings and break them down into smaller questions.**\n - **If a question cannot be broken down, present the question as the output.**\n - **Stick to the context of the questions and optimize the subquestions to minimize their number.**\n\n2. **Analyze each subquestion and categorize it as \"ClinicalTrials,\" \"DailyMed,\" or \"Internet\" based on the provided definitions:**\n - **\"ClinicalTrials\" refers exclusively to data connected with clinical trials, such as study particulars, conditions, interventions, eligibility criteria, locations, sponsors/collaborators, outcome measures, and efficacy/safety outcomes.**\n - **\"DailyMed\" solely includes data about drug names, approved uses, dosage, administration, side effects, and contraindications.**\n - **\"Internet\" encompasses data unrelated to both \"ClinicalTrials\" and \"DailyMed\"; this can include tasks performed by the Language Model.**\n\n3. **Evaluate each subquestion and determine if it contains a date range with a start and end date or output \"NA\" following the provided guidelines:**\n - **If terms like \"current,\" \"recent,\" \"latest,\" or \"recently\" are used, the date filter should span the last year from the current date, which is 2023-08-11.**\n - **If an explicit date is mentioned, use the provided start and end dates for the date range.**\n - **Represent the date range in the format [\"YYYY-MM-DD\", \"YYYY-MM-DD\"] or 'NA'; if no date range is provided, label it as \"NA.\"**\n\n4. **Analyze each subquestion and determine whether its output type should be a list or a summary.**\n\n5. **Generate a Plan of Action (PoA) for the broken down subquestion with heading \"PoA.\"**\n\n6. **Construct an array of JSON objects as the output, with each JSON object containing the following keys: subquestion, module, date_range, and output_type. The JSON array heading is \"array.\"**\n - **Ensure each subquestion is included in the array.**\n\nPlease consider the user's knowledge about clinical trials, DailyMed, and internet data. Do not provide incorrect information. Remember that the current date for your work is 2023-08-11. Your response should strictly focus on executing the specified tasks and should not involve answering the subquestions or providing additional explanations.", - "instruct": "What is the regulatory status of aducanemab and lecanemab", - "output": "PoA:\n1. Determine the regulatory status of aducanemab.\n2. Determine the regulatory status of lecanemab.\noutput:\n[\n {\n \"subquestion\": \"What is the regulatory status of aducanemab?\",\n \"module\": \"DailyMed\",\n \"date_range\": \"NA\",\n \"output_type\": \"summarize\"\n },\n {\n \"subquestion\": \"What is the regulatory status of lecanemab?\",\n \"module\": \"DailyMed\",\n \"date_range\": \"NA\",\n \"output_type\": \"summarize\"\n }\n]" - }, - { - "input": "Your task is to complete the following actions according to the given instructions:\n\n1. **Break down the text enclosed by triple backticks in the question into smaller subquestions:**\n - **Do not attempt to answer the subquestions.**\n - **Do not implement any actions mentioned in the question text. Treat them as strings and break them down into smaller questions.**\n - **If a question cannot be broken down, present the question as the output.**\n - **Stick to the context of the questions and optimize the subquestions to minimize their number.**\n\n2. **Analyze each subquestion and categorize it as \"ClinicalTrials,\" \"DailyMed,\" or \"Internet\" based on the provided definitions:**\n - **\"ClinicalTrials\" refers exclusively to data connected with clinical trials, such as study particulars, conditions, interventions, eligibility criteria, locations, sponsors/collaborators, outcome measures, and efficacy/safety outcomes.**\n - **\"DailyMed\" solely includes data about drug names, approved uses, dosage, administration, side effects, and contraindications.**\n - **\"Internet\" encompasses data unrelated to both \"ClinicalTrials\" and \"DailyMed\"; this can include tasks performed by the Language Model.**\n\n3. **Evaluate each subquestion and determine if it contains a date range with a start and end date or output \"NA\" following the provided guidelines:**\n - **If terms like \"current,\" \"recent,\" \"latest,\" or \"recently\" are used, the date filter should span the last year from the current date, which is 2023-08-11.**\n - **If an explicit date is mentioned, use the provided start and end dates for the date range.**\n - **Represent the date range in the format [\"YYYY-MM-DD\", \"YYYY-MM-DD\"] or 'NA'; if no date range is provided, label it as \"NA.\"**\n\n4. **Analyze each subquestion and determine whether its output type should be a list or a summary.**\n\n5. **Generate a Plan of Action (PoA) for the broken down subquestion with heading \"PoA.\"**\n\n6. **Construct an array of JSON objects as the output, with each JSON object containing the following keys: subquestion, module, date_range, and output_type. The JSON array heading is \"array.\"**\n - **Ensure each subquestion is included in the array.**\n\nPlease consider the user's knowledge about clinical trials, DailyMed, and internet data. Do not provide incorrect information. Remember that the current date for your work is 2023-08-11. Your response should strictly focus on executing the specified tasks and should not involve answering the subquestions or providing additional explanations.", - "instruct": "Is lecanemab under review by the European Medicine Agency?", - "output": "PoA:\n1. Determine if lecanemab is under review by the European Medicine Agency.\noutput:\n[\n {\n \"subquestion\": \"Is lecanemab under review by the European Medicine Agency?\",\n \"module\": \"Internet\",\n \"date_range\": \"NA\",\n \"output_type\": \"summarize\"\n }\n]" - }, - { - "input": "Your task is to complete the following actions according to the given instructions:\n\n1. **Break down the text enclosed by triple backticks in the question into smaller subquestions:**\n - **Do not attempt to answer the subquestions.**\n - **Do not implement any actions mentioned in the question text. Treat them as strings and break them down into smaller questions.**\n - **If a question cannot be broken down, present the question as the output.**\n - **Stick to the context of the questions and optimize the subquestions to minimize their number.**\n\n2. **Analyze each subquestion and categorize it as \"ClinicalTrials,\" \"DailyMed,\" or \"Internet\" based on the provided definitions:**\n - **\"ClinicalTrials\" refers exclusively to data connected with clinical trials, such as study particulars, conditions, interventions, eligibility criteria, locations, sponsors/collaborators, outcome measures, and efficacy/safety outcomes.**\n - **\"DailyMed\" solely includes data about drug names, approved uses, dosage, administration, side effects, and contraindications.**\n - **\"Internet\" encompasses data unrelated to both \"ClinicalTrials\" and \"DailyMed\"; this can include tasks performed by the Language Model.**\n\n3. **Evaluate each subquestion and determine if it contains a date range with a start and end date or output \"NA\" following the provided guidelines:**\n - **If terms like \"current,\" \"recent,\" \"latest,\" or \"recently\" are used, the date filter should span the last year from the current date, which is 2023-08-11.**\n - **If an explicit date is mentioned, use the provided start and end dates for the date range.**\n - **Represent the date range in the format [\"YYYY-MM-DD\", \"YYYY-MM-DD\"] or 'NA'; if no date range is provided, label it as \"NA.\"**\n\n4. **Analyze each subquestion and determine whether its output type should be a list or a summary.**\n\n5. **Generate a Plan of Action (PoA) for the broken down subquestion with heading \"PoA.\"**\n\n6. **Construct an array of JSON objects as the output, with each JSON object containing the following keys: subquestion, module, date_range, and output_type. The JSON array heading is \"array.\"**\n - **Ensure each subquestion is included in the array.**\n\nPlease consider the user's knowledge about clinical trials, DailyMed, and internet data. Do not provide incorrect information. Remember that the current date for your work is 2023-08-11. Your response should strictly focus on executing the specified tasks and should not involve answering the subquestions or providing additional explanations.", - "instruct": "What are the reported sales for Keytruda in 2023 and its forecast?", - "output": "PoA:\n1. Identify the reported sales for Keytruda in 2023.\n2. Research the forecast for Keytruda sales in 2023.\noutput:\n[\n {\n \"subquestion\": \"What are the reported sales for Keytruda in 2023?\",\n \"module\": \"Internet\",\n \"date_range\": [\"2023-01-01\", \"2023-12-31\"],\n \"output_type\": \"List\"\n },\n {\n \"subquestion\": \"What is the forecast for Keytruda sales in 2023?\",\n \"module\": \"Internet\",\n \"date_range\": [\"2023-01-01\", \"2023-12-31\"],\n \"output_type\": \"summarize\"\n }\n]" - }, - { - "input": "Your task is to complete the following actions according to the given instructions:\n\n1. **Break down the text enclosed by triple backticks in the question into smaller subquestions:**\n - **Do not attempt to answer the subquestions.**\n - **Do not implement any actions mentioned in the question text. Treat them as strings and break them down into smaller questions.**\n - **If a question cannot be broken down, present the question as the output.**\n - **Stick to the context of the questions and optimize the subquestions to minimize their number.**\n\n2. **Analyze each subquestion and categorize it as \"ClinicalTrials,\" \"DailyMed,\" or \"Internet\" based on the provided definitions:**\n - **\"ClinicalTrials\" refers exclusively to data connected with clinical trials, such as study particulars, conditions, interventions, eligibility criteria, locations, sponsors/collaborators, outcome measures, and efficacy/safety outcomes.**\n - **\"DailyMed\" solely includes data about drug names, approved uses, dosage, administration, side effects, and contraindications.**\n - **\"Internet\" encompasses data unrelated to both \"ClinicalTrials\" and \"DailyMed\"; this can include tasks performed by the Language Model.**\n\n3. **Evaluate each subquestion and determine if it contains a date range with a start and end date or output \"NA\" following the provided guidelines:**\n - **If terms like \"current,\" \"recent,\" \"latest,\" or \"recently\" are used, the date filter should span the last year from the current date, which is 2023-08-11.**\n - **If an explicit date is mentioned, use the provided start and end dates for the date range.**\n - **Represent the date range in the format [\"YYYY-MM-DD\", \"YYYY-MM-DD\"] or 'NA'; if no date range is provided, label it as \"NA.\"**\n\n4. **Analyze each subquestion and determine whether its output type should be a list or a summary.**\n\n5. **Generate a Plan of Action (PoA) for the broken down subquestion with heading \"PoA.\"**\n\n6. **Construct an array of JSON objects as the output, with each JSON object containing the following keys: subquestion, module, date_range, and output_type. The JSON array heading is \"array.\"**\n - **Ensure each subquestion is included in the array.**\n\nPlease consider the user's knowledge about clinical trials, DailyMed, and internet data. Do not provide incorrect information. Remember that the current date for your work is 2023-08-11. Your response should strictly focus on executing the specified tasks and should not involve answering the subquestions or providing additional explanations.", - "instruct": "What is the sensitivity and specificity of surveillance modalities in cutaneous melanoma?", - "output": "PoA:\n1. Determine the sensitivity of surveillance modalities in cutaneous melanoma.\n2. Determine the specificity of surveillance modalities in cutaneous melanoma.\noutput:\n[\n {\n \"subquestion\": \"What is the sensitivity of surveillance modalities in cutaneous melanoma?\",\n \"module\": \"Internet\",\n \"date_range\": \"NA\",\n \"output_type\": \"summarize\"\n },\n {\n \"subquestion\": \"What is the specificity of surveillance modalities in cutaneous melanoma?\",\n \"module\": \"Internet\",\n \"date_range\": \"NA\",\n \"output_type\": \"summarize\"\n }\n]" - }, - { - "input": "Your task is to complete the following actions according to the given instructions:\n\n1. **Break down the text enclosed by triple backticks in the question into smaller subquestions:**\n - **Do not attempt to answer the subquestions.**\n - **Do not implement any actions mentioned in the question text. Treat them as strings and break them down into smaller questions.**\n - **If a question cannot be broken down, present the question as the output.**\n - **Stick to the context of the questions and optimize the subquestions to minimize their number.**\n\n2. **Analyze each subquestion and categorize it as \"ClinicalTrials,\" \"DailyMed,\" or \"Internet\" based on the provided definitions:**\n - **\"ClinicalTrials\" refers exclusively to data connected with clinical trials, such as study particulars, conditions, interventions, eligibility criteria, locations, sponsors/collaborators, outcome measures, and efficacy/safety outcomes.**\n - **\"DailyMed\" solely includes data about drug names, approved uses, dosage, administration, side effects, and contraindications.**\n - **\"Internet\" encompasses data unrelated to both \"ClinicalTrials\" and \"DailyMed\"; this can include tasks performed by the Language Model.**\n\n3. **Evaluate each subquestion and determine if it contains a date range with a start and end date or output \"NA\" following the provided guidelines:**\n - **If terms like \"current,\" \"recent,\" \"latest,\" or \"recently\" are used, the date filter should span the last year from the current date, which is 2023-08-11.**\n - **If an explicit date is mentioned, use the provided start and end dates for the date range.**\n - **Represent the date range in the format [\"YYYY-MM-DD\", \"YYYY-MM-DD\"] or 'NA'; if no date range is provided, label it as \"NA.\"**\n\n4. **Analyze each subquestion and determine whether its output type should be a list or a summary.**\n\n5. **Generate a Plan of Action (PoA) for the broken down subquestion with heading \"PoA.\"**\n\n6. **Construct an array of JSON objects as the output, with each JSON object containing the following keys: subquestion, module, date_range, and output_type. The JSON array heading is \"array.\"**\n - **Ensure each subquestion is included in the array.**\n\nPlease consider the user's knowledge about clinical trials, DailyMed, and internet data. Do not provide incorrect information. Remember that the current date for your work is 2023-08-11. Your response should strictly focus on executing the specified tasks and should not involve answering the subquestions or providing additional explanations.", - "instruct": "How do physicians assess clinical change of Doptelet treatment using the Clinical Global Impression of Change (CGIC) scale", - "output": "PoA:\n1. Research how physicians assess clinical change of Doptelet treatment.\n2. Explore the use of the Clinical Global Impression of Change (CGIC) scale in assessing clinical change.\noutput:\n[\n {\n \"subquestion\": \"What is the Clinical Global Impression of Change (CGIC) scale?\",\n \"module\": \"Internet\",\n \"date_range\": \"NA\",\n \"output_type\": \"summarize\"\n },\n {\n \"subquestion\": \"How do physicians assess clinical change of Doptelet treatment using the CGIC scale?\",\n \"module\": \"Internet\",\n \"date_range\": \"NA\",\n \"output_type\": \"summarize\"\n }\n]" - }, - { - "input": "Your task is to complete the following actions according to the given instructions:\n\n1. **Break down the text enclosed by triple backticks in the question into smaller subquestions:**\n - **Do not attempt to answer the subquestions.**\n - **Do not implement any actions mentioned in the question text. Treat them as strings and break them down into smaller questions.**\n - **If a question cannot be broken down, present the question as the output.**\n - **Stick to the context of the questions and optimize the subquestions to minimize their number.**\n\n2. **Analyze each subquestion and categorize it as \"ClinicalTrials,\" \"DailyMed,\" or \"Internet\" based on the provided definitions:**\n - **\"ClinicalTrials\" refers exclusively to data connected with clinical trials, such as study particulars, conditions, interventions, eligibility criteria, locations, sponsors/collaborators, outcome measures, and efficacy/safety outcomes.**\n - **\"DailyMed\" solely includes data about drug names, approved uses, dosage, administration, side effects, and contraindications.**\n - **\"Internet\" encompasses data unrelated to both \"ClinicalTrials\" and \"DailyMed\"; this can include tasks performed by the Language Model.**\n\n3. **Evaluate each subquestion and determine if it contains a date range with a start and end date or output \"NA\" following the provided guidelines:**\n - **If terms like \"current,\" \"recent,\" \"latest,\" or \"recently\" are used, the date filter should span the last year from the current date, which is 2023-08-11.**\n - **If an explicit date is mentioned, use the provided start and end dates for the date range.**\n - **Represent the date range in the format [\"YYYY-MM-DD\", \"YYYY-MM-DD\"] or 'NA'; if no date range is provided, label it as \"NA.\"**\n\n4. **Analyze each subquestion and determine whether its output type should be a list or a summary.**\n\n5. **Generate a Plan of Action (PoA) for the broken down subquestion with heading \"PoA.\"**\n\n6. **Construct an array of JSON objects as the output, with each JSON object containing the following keys: subquestion, module, date_range, and output_type. The JSON array heading is \"array.\"**\n - **Ensure each subquestion is included in the array.**\n\nPlease consider the user's knowledge about clinical trials, DailyMed, and internet data. Do not provide incorrect information. Remember that the current date for your work is 2023-08-11. Your response should strictly focus on executing the specified tasks and should not involve answering the subquestions or providing additional explanations.", - "instruct": "Based on Roche's 2023 Q1 earnings presentation, what diseases is Roche studying Enspryng in", - "output": "PoA:\n1. Identify Roche's 2023 Q1 earnings presentation.\n2. Determine the diseases that Roche is studying Enspryng in based on the presentation.\noutput:\n[\n {\n \"subquestion\": \"Based on Roche's 2023 Q1 earnings presentation, what diseases is Roche studying Enspryng in?\",\n \"module\": \"Internet\",\n \"date_range\": [\"2023-01-01\", \"2023-03-31\"],\n \"output_type\": \"summarize\"\n }\n]" - }, - { - "input": "Your task is to complete the following actions according to the given instructions:\n\n1. **Break down the text enclosed by triple backticks in the question into smaller subquestions:**\n - **Do not attempt to answer the subquestions.**\n - **Do not implement any actions mentioned in the question text. Treat them as strings and break them down into smaller questions.**\n - **If a question cannot be broken down, present the question as the output.**\n - **Stick to the context of the questions and optimize the subquestions to minimize their number.**\n\n2. **Analyze each subquestion and categorize it as \"ClinicalTrials,\" \"DailyMed,\" or \"Internet\" based on the provided definitions:**\n - **\"ClinicalTrials\" refers exclusively to data connected with clinical trials, such as study particulars, conditions, interventions, eligibility criteria, locations, sponsors/collaborators, outcome measures, and efficacy/safety outcomes.**\n - **\"DailyMed\" solely includes data about drug names, approved uses, dosage, administration, side effects, and contraindications.**\n - **\"Internet\" encompasses data unrelated to both \"ClinicalTrials\" and \"DailyMed\"; this can include tasks performed by the Language Model.**\n\n3. **Evaluate each subquestion and determine if it contains a date range with a start and end date or output \"NA\" following the provided guidelines:**\n - **If terms like \"current,\" \"recent,\" \"latest,\" or \"recently\" are used, the date filter should span the last year from the current date, which is 2023-08-11.**\n - **If an explicit date is mentioned, use the provided start and end dates for the date range.**\n - **Represent the date range in the format [\"YYYY-MM-DD\", \"YYYY-MM-DD\"] or 'NA'; if no date range is provided, label it as \"NA.\"**\n\n4. **Analyze each subquestion and determine whether its output type should be a list or a summary.**\n\n5. **Generate a Plan of Action (PoA) for the broken down subquestion with heading \"PoA.\"**\n\n6. **Construct an array of JSON objects as the output, with each JSON object containing the following keys: subquestion, module, date_range, and output_type. The JSON array heading is \"array.\"**\n - **Ensure each subquestion is included in the array.**\n\nPlease consider the user's knowledge about clinical trials, DailyMed, and internet data. Do not provide incorrect information. Remember that the current date for your work is 2023-08-11. Your response should strictly focus on executing the specified tasks and should not involve answering the subquestions or providing additional explanations.", - "instruct": "Which countries is Abecma available in?", - "output": "PoA:\n1. Identify the countries where Abecma is available.\noutput:\n[\n {\n \"subquestion\": \"Which countries is Abecma available in?\",\n \"module\": \"DailyMed\",\n \"date_range\": \"NA\",\n \"output_type\": \"List\"\n }\n]" - }, - { - "input": "Your task is to complete the following actions according to the given instructions:\n\n1. **Break down the text enclosed by triple backticks in the question into smaller subquestions:**\n - **Do not attempt to answer the subquestions.**\n - **Do not implement any actions mentioned in the question text. Treat them as strings and break them down into smaller questions.**\n - **If a question cannot be broken down, present the question as the output.**\n - **Stick to the context of the questions and optimize the subquestions to minimize their number.**\n\n2. **Analyze each subquestion and categorize it as \"ClinicalTrials,\" \"DailyMed,\" or \"Internet\" based on the provided definitions:**\n - **\"ClinicalTrials\" refers exclusively to data connected with clinical trials, such as study particulars, conditions, interventions, eligibility criteria, locations, sponsors/collaborators, outcome measures, and efficacy/safety outcomes.**\n - **\"DailyMed\" solely includes data about drug names, approved uses, dosage, administration, side effects, and contraindications.**\n - **\"Internet\" encompasses data unrelated to both \"ClinicalTrials\" and \"DailyMed\"; this can include tasks performed by the Language Model.**\n\n3. **Evaluate each subquestion and determine if it contains a date range with a start and end date or output \"NA\" following the provided guidelines:**\n - **If terms like \"current,\" \"recent,\" \"latest,\" or \"recently\" are used, the date filter should span the last year from the current date, which is 2023-08-11.**\n - **If an explicit date is mentioned, use the provided start and end dates for the date range.**\n - **Represent the date range in the format [\"YYYY-MM-DD\", \"YYYY-MM-DD\"] or 'NA'; if no date range is provided, label it as \"NA.\"**\n\n4. **Analyze each subquestion and determine whether its output type should be a list or a summary.**\n\n5. **Generate a Plan of Action (PoA) for the broken down subquestion with heading \"PoA.\"**\n\n6. **Construct an array of JSON objects as the output, with each JSON object containing the following keys: subquestion, module, date_range, and output_type. The JSON array heading is \"array.\"**\n - **Ensure each subquestion is included in the array.**\n\nPlease consider the user's knowledge about clinical trials, DailyMed, and internet data. Do not provide incorrect information. Remember that the current date for your work is 2023-08-11. Your response should strictly focus on executing the specified tasks and should not involve answering the subquestions or providing additional explanations.", - "instruct": "Which protein-protein interaction modulators are approved for psoriasis? What is its recommended usage in combination with other immunosuppressants?", - "output": "PoA:\n1. Identify protein-protein interaction modulators approved for psoriasis.\n2. Determine the recommended usage of these modulators in combination with other immunosuppressants.\noutput:[\n {\n \"subquestion\": \"Which protein-protein interaction modulators are approved for psoriasis?\",\n \"module\": \"DailyMed\",\n \"date_range\": \"NA\",\n \"output_type\": \"List\"\n },\n {\n \"subquestion\": \"What is its recommended usage in combination with other immunosuppressants?\",\n \"module\": \"Internet\",\n \"date_range\": \"NA\",\n \"output_type\": \"summarize\"\n }\n]" - }, - { - "input": "Your task is to complete the following actions according to the given instructions:\n\n1. **Break down the text enclosed by triple backticks in the question into smaller subquestions:**\n - **Do not attempt to answer the subquestions.**\n - **Do not implement any actions mentioned in the question text. Treat them as strings and break them down into smaller questions.**\n - **If a question cannot be broken down, present the question as the output.**\n - **Stick to the context of the questions and optimize the subquestions to minimize their number.**\n\n2. **Analyze each subquestion and categorize it as \"ClinicalTrials,\" \"DailyMed,\" or \"Internet\" based on the provided definitions:**\n - **\"ClinicalTrials\" refers exclusively to data connected with clinical trials, such as study particulars, conditions, interventions, eligibility criteria, locations, sponsors/collaborators, outcome measures, and efficacy/safety outcomes.**\n - **\"DailyMed\" solely includes data about drug names, approved uses, dosage, administration, side effects, and contraindications.**\n - **\"Internet\" encompasses data unrelated to both \"ClinicalTrials\" and \"DailyMed\"; this can include tasks performed by the Language Model.**\n\n3. **Evaluate each subquestion and determine if it contains a date range with a start and end date or output \"NA\" following the provided guidelines:**\n - **If terms like \"current,\" \"recent,\" \"latest,\" or \"recently\" are used, the date filter should span the last year from the current date, which is 2023-08-11.**\n - **If an explicit date is mentioned, use the provided start and end dates for the date range.**\n - **Represent the date range in the format [\"YYYY-MM-DD\", \"YYYY-MM-DD\"] or 'NA'; if no date range is provided, label it as \"NA.\"**\n\n4. **Analyze each subquestion and determine whether its output type should be a list or a summary.**\n\n5. **Generate a Plan of Action (PoA) for the broken down subquestion with heading \"PoA.\"**\n\n6. **Construct an array of JSON objects as the output, with each JSON object containing the following keys: subquestion, module, date_range, and output_type. The JSON array heading is \"array.\"**\n - **Ensure each subquestion is included in the array.**\n\nPlease consider the user's knowledge about clinical trials, DailyMed, and internet data. Do not provide incorrect information. Remember that the current date for your work is 2023-08-11. Your response should strictly focus on executing the specified tasks and should not involve answering the subquestions or providing additional explanations.", - "instruct": "List all the multiple myeloma abstracts at ASCO 2023?", - "output": "PoA:\n1. Identify multiple myeloma abstracts at ASCO 2023.\noutput:[\n {\n \"subquestion\": \"List all the multiple myeloma abstracts at ASCO 2023?\",\n \"module\": \"Internet\",\n \"date_range\": \"NA\",\n \"output_type\": \"List\"\n }\n]" - }, - { - "input": "Your task is to complete the following actions according to the given instructions:\n\n1. **Break down the text enclosed by triple backticks in the question into smaller subquestions:**\n - **Do not attempt to answer the subquestions.**\n - **Do not implement any actions mentioned in the question text. Treat them as strings and break them down into smaller questions.**\n - **If a question cannot be broken down, present the question as the output.**\n - **Stick to the context of the questions and optimize the subquestions to minimize their number.**\n\n2. **Analyze each subquestion and categorize it as \"ClinicalTrials,\" \"DailyMed,\" or \"Internet\" based on the provided definitions:**\n - **\"ClinicalTrials\" refers exclusively to data connected with clinical trials, such as study particulars, conditions, interventions, eligibility criteria, locations, sponsors/collaborators, outcome measures, and efficacy/safety outcomes.**\n - **\"DailyMed\" solely includes data about drug names, approved uses, dosage, administration, side effects, and contraindications.**\n - **\"Internet\" encompasses data unrelated to both \"ClinicalTrials\" and \"DailyMed\"; this can include tasks performed by the Language Model.**\n\n3. **Evaluate each subquestion and determine if it contains a date range with a start and end date or output \"NA\" following the provided guidelines:**\n - **If terms like \"current,\" \"recent,\" \"latest,\" or \"recently\" are used, the date filter should span the last year from the current date, which is 2023-08-11.**\n - **If an explicit date is mentioned, use the provided start and end dates for the date range.**\n - **Represent the date range in the format [\"YYYY-MM-DD\", \"YYYY-MM-DD\"] or 'NA'; if no date range is provided, label it as \"NA.\"**\n\n4. **Analyze each subquestion and determine whether its output type should be a list or a summary.**\n\n5. **Generate a Plan of Action (PoA) for the broken down subquestion with heading \"PoA.\"**\n\n6. **Construct an array of JSON objects as the output, with each JSON object containing the following keys: subquestion, module, date_range, and output_type. The JSON array heading is \"array.\"**\n - **Ensure each subquestion is included in the array.**\n\nPlease consider the user's knowledge about clinical trials, DailyMed, and internet data. Do not provide incorrect information. Remember that the current date for your work is 2023-08-11. Your response should strictly focus on executing the specified tasks and should not involve answering the subquestions or providing additional explanations.", - "instruct": "Are there any new drugs being investigated in the US for Gaucher Disease by an industry sponsor?", - "output": "PoA:\n1. Identify new drugs being investigated in the US for Gaucher Disease.\noutput:[\n {\n \"subquestion\": \"Are there any new drugs being investigated in the US for Gaucher Disease by an industry sponsor?\",\n \"module\": \"ClinicalTrials\",\n \"date_range\": \"NA\",\n \"output_type\": \"List\"\n }\n]" - }, - { - "input": "Your task is to complete the following actions according to the given instructions:\n\n1. **Break down the text enclosed by triple backticks in the question into smaller subquestions:**\n - **Do not attempt to answer the subquestions.**\n - **Do not implement any actions mentioned in the question text. Treat them as strings and break them down into smaller questions.**\n - **If a question cannot be broken down, present the question as the output.**\n - **Stick to the context of the questions and optimize the subquestions to minimize their number.**\n\n2. **Analyze each subquestion and categorize it as \"ClinicalTrials,\" \"DailyMed,\" or \"Internet\" based on the provided definitions:**\n - **\"ClinicalTrials\" refers exclusively to data connected with clinical trials, such as study particulars, conditions, interventions, eligibility criteria, locations, sponsors/collaborators, outcome measures, and efficacy/safety outcomes.**\n - **\"DailyMed\" solely includes data about drug names, approved uses, dosage, administration, side effects, and contraindications.**\n - **\"Internet\" encompasses data unrelated to both \"ClinicalTrials\" and \"DailyMed\"; this can include tasks performed by the Language Model.**\n\n3. **Evaluate each subquestion and determine if it contains a date range with a start and end date or output \"NA\" following the provided guidelines:**\n - **If terms like \"current,\" \"recent,\" \"latest,\" or \"recently\" are used, the date filter should span the last year from the current date, which is 2023-08-11.**\n - **If an explicit date is mentioned, use the provided start and end dates for the date range.**\n - **Represent the date range in the format [\"YYYY-MM-DD\", \"YYYY-MM-DD\"] or 'NA'; if no date range is provided, label it as \"NA.\"**\n\n4. **Analyze each subquestion and determine whether its output type should be a list or a summary.**\n\n5. **Generate a Plan of Action (PoA) for the broken down subquestion with heading \"PoA.\"**\n\n6. **Construct an array of JSON objects as the output, with each JSON object containing the following keys: subquestion, module, date_range, and output_type. The JSON array heading is \"array.\"**\n - **Ensure each subquestion is included in the array.**\n\nPlease consider the user's knowledge about clinical trials, DailyMed, and internet data. Do not provide incorrect information. Remember that the current date for your work is 2023-08-11. Your response should strictly focus on executing the specified tasks and should not involve answering the subquestions or providing additional explanations.", - "instruct": "Who are the main competitors for osimertinib for treatment of non-small cell lung cancer?", - "output": "PoA:\n1. Identify the main competitors for osimertinib in the treatment of non-small cell lung cancer.\noutput:[\n {\n \"subquestion\": \"Who are the main competitors for osimertinib for treatment of non-small cell lung cancer?\",\n \"module\": \"Internet\",\n \"date_range\": \"NA\",\n \"output_type\": \"List\"\n }\n]" - }, - { - "input": "Your task is to complete the following actions according to the given instructions:\n\n1. **Break down the text enclosed by triple backticks in the question into smaller subquestions:**\n - **Do not attempt to answer the subquestions.**\n - **Do not implement any actions mentioned in the question text. Treat them as strings and break them down into smaller questions.**\n - **If a question cannot be broken down, present the question as the output.**\n - **Stick to the context of the questions and optimize the subquestions to minimize their number.**\n\n2. **Analyze each subquestion and categorize it as \"ClinicalTrials,\" \"DailyMed,\" or \"Internet\" based on the provided definitions:**\n - **\"ClinicalTrials\" refers exclusively to data connected with clinical trials, such as study particulars, conditions, interventions, eligibility criteria, locations, sponsors/collaborators, outcome measures, and efficacy/safety outcomes.**\n - **\"DailyMed\" solely includes data about drug names, approved uses, dosage, administration, side effects, and contraindications.**\n - **\"Internet\" encompasses data unrelated to both \"ClinicalTrials\" and \"DailyMed\"; this can include tasks performed by the Language Model.**\n\n3. **Evaluate each subquestion and determine if it contains a date range with a start and end date or output \"NA\" following the provided guidelines:**\n - **If terms like \"current,\" \"recent,\" \"latest,\" or \"recently\" are used, the date filter should span the last year from the current date, which is 2023-08-11.**\n - **If an explicit date is mentioned, use the provided start and end dates for the date range.**\n - **Represent the date range in the format [\"YYYY-MM-DD\", \"YYYY-MM-DD\"] or 'NA'; if no date range is provided, label it as \"NA.\"**\n\n4. **Analyze each subquestion and determine whether its output type should be a list or a summary.**\n\n5. **Generate a Plan of Action (PoA) for the broken down subquestion with heading \"PoA.\"**\n\n6. **Construct an array of JSON objects as the output, with each JSON object containing the following keys: subquestion, module, date_range, and output_type. The JSON array heading is \"array.\"**\n - **Ensure each subquestion is included in the array.**\n\nPlease consider the user's knowledge about clinical trials, DailyMed, and internet data. Do not provide incorrect information. Remember that the current date for your work is 2023-08-11. Your response should strictly focus on executing the specified tasks and should not involve answering the subquestions or providing additional explanations.", - "instruct": "Compare LUMAKRAS and KRAZATI across indications, Q1 2023 sales and active phase 3 trials.", - "output": "PoA:\n1. Compare the indications of LUMAKRAS and KRAZATI.\n2. Analyze the Q1 2023 sales of LUMAKRAS and KRAZATI.\n3. Determine the number of active phase 3 trials for LUMAKRAS and KRAZATI.\noutput:[\n {\n \"subquestion\": \"Compare the indications of LUMAKRAS and KRAZATI.\",\n \"module\": \"DailyMed\",\n \"date_range\": \"NA\",\n \"output_type\": \"summarize\"\n },\n {\n \"subquestion\": \"Analyze the Q1 2023 sales of LUMAKRAS and KRAZATI.\",\n \"module\": \"Internet\",\n \"date_range\": [\"2023-01-01\", \"2023-03-31\"],\n \"output_type\": \"summarize\"\n },\n {\n \"subquestion\": \"Determine the number of active phase 3 trials for LUMAKRAS and KRAZATI.\",\n \"module\": \"ClinicalTrials\",\n \"date_range\": \"NA\",\n \"output_type\": \"List\"\n }\n]" - }, - { - "input": "Your task is to complete the following actions according to the given instructions:\n\n1. **Break down the text enclosed by triple backticks in the question into smaller subquestions:**\n - **Do not attempt to answer the subquestions.**\n - **Do not implement any actions mentioned in the question text. Treat them as strings and break them down into smaller questions.**\n - **If a question cannot be broken down, present the question as the output.**\n - **Stick to the context of the questions and optimize the subquestions to minimize their number.**\n\n2. **Analyze each subquestion and categorize it as \"ClinicalTrials,\" \"DailyMed,\" or \"Internet\" based on the provided definitions:**\n - **\"ClinicalTrials\" refers exclusively to data connected with clinical trials, such as study particulars, conditions, interventions, eligibility criteria, locations, sponsors/collaborators, outcome measures, and efficacy/safety outcomes.**\n - **\"DailyMed\" solely includes data about drug names, approved uses, dosage, administration, side effects, and contraindications.**\n - **\"Internet\" encompasses data unrelated to both \"ClinicalTrials\" and \"DailyMed\"; this can include tasks performed by the Language Model.**\n\n3. **Evaluate each subquestion and determine if it contains a date range with a start and end date or output \"NA\" following the provided guidelines:**\n - **If terms like \"current,\" \"recent,\" \"latest,\" or \"recently\" are used, the date filter should span the last year from the current date, which is 2023-08-11.**\n - **If an explicit date is mentioned, use the provided start and end dates for the date range.**\n - **Represent the date range in the format [\"YYYY-MM-DD\", \"YYYY-MM-DD\"] or 'NA'; if no date range is provided, label it as \"NA.\"**\n\n4. **Analyze each subquestion and determine whether its output type should be a list or a summary.**\n\n5. **Generate a Plan of Action (PoA) for the broken down subquestion with heading \"PoA.\"**\n\n6. **Construct an array of JSON objects as the output, with each JSON object containing the following keys: subquestion, module, date_range, and output_type. The JSON array heading is \"array.\"**\n - **Ensure each subquestion is included in the array.**\n\nPlease consider the user's knowledge about clinical trials, DailyMed, and internet data. Do not provide incorrect information. Remember that the current date for your work is 2023-08-11. Your response should strictly focus on executing the specified tasks and should not involve answering the subquestions or providing additional explanations.", - "instruct": "Please create a table summarizing all drugs in active development for hidradenitis suppurativa with a column listing the primary completion date for each trial", - "output": "PoA:\n1. Identify drugs in active development for hidradenitis suppurativa.\n2. Create a table summarizing the drugs with a column for the primary completion date of each trial.\noutput: [\n {\n \"subquestion\": \"What drugs are in active development for hidradenitis suppurativa?\",\n \"module\": \"ClinicalTrials\",\n \"date_range\": \"NA\",\n \"output_type\": \"List\"\n },\n {\n \"subquestion\": \"Create a table summarizing the drugs in active development for hidradenitis suppurativa with a column for the primary completion date of each trial.\",\n \"module\": \"ClinicalTrials\",\n \"date_range\": \"NA\",\n \"output_type\": \"summarize\"\n }\n ]" - }, - { - "input": "Your task is to complete the following actions according to the given instructions:\n\n1. **Break down the text enclosed by triple backticks in the question into smaller subquestions:**\n - **Do not attempt to answer the subquestions.**\n - **Do not implement any actions mentioned in the question text. Treat them as strings and break them down into smaller questions.**\n - **If a question cannot be broken down, present the question as the output.**\n - **Stick to the context of the questions and optimize the subquestions to minimize their number.**\n\n2. **Analyze each subquestion and categorize it as \"ClinicalTrials,\" \"DailyMed,\" or \"Internet\" based on the provided definitions:**\n - **\"ClinicalTrials\" refers exclusively to data connected with clinical trials, such as study particulars, conditions, interventions, eligibility criteria, locations, sponsors/collaborators, outcome measures, and efficacy/safety outcomes.**\n - **\"DailyMed\" solely includes data about drug names, approved uses, dosage, administration, side effects, and contraindications.**\n - **\"Internet\" encompasses data unrelated to both \"ClinicalTrials\" and \"DailyMed\"; this can include tasks performed by the Language Model.**\n\n3. **Evaluate each subquestion and determine if it contains a date range with a start and end date or output \"NA\" following the provided guidelines:**\n - **If terms like \"current,\" \"recent,\" \"latest,\" or \"recently\" are used, the date filter should span the last year from the current date, which is 2023-08-11.**\n - **If an explicit date is mentioned, use the provided start and end dates for the date range.**\n - **Represent the date range in the format [\"YYYY-MM-DD\", \"YYYY-MM-DD\"] or 'NA'; if no date range is provided, label it as \"NA.\"**\n\n4. **Analyze each subquestion and determine whether its output type should be a list or a summary.**\n\n5. **Generate a Plan of Action (PoA) for the broken down subquestion with heading \"PoA.\"**\n\n6. **Construct an array of JSON objects as the output, with each JSON object containing the following keys: subquestion, module, date_range, and output_type. The JSON array heading is \"array.\"**\n - **Ensure each subquestion is included in the array.**\n\nPlease consider the user's knowledge about clinical trials, DailyMed, and internet data. Do not provide incorrect information. Remember that the current date for your work is 2023-08-11. Your response should strictly focus on executing the specified tasks and should not involve answering the subquestions or providing additional explanations.", - "instruct": "Can you summarize the ongoing trials for belzutifan in table format. I'd like it to include key parameters like the trial number, combination agents, indication / line of therapy, primary endpoint, phase of study, primary completion date.", - "output": "PoA:\n1. Identify ongoing trials for belzutifan.\n2. Extract key parameters such as trial number, combination agents, indication/line of therapy, primary endpoint, phase of study, and primary completion date.\n3. Summarize the information in table format.\noutput: [\n {\n \"subquestion\": \"What are the ongoing trials for belzutifan?\",\n \"module\": \"ClinicalTrials\",\n \"date_range\": \"NA\",\n \"output_type\": \"List\"\n },\n {\n \"subquestion\": \"Extract trial number, combination agents, indication/line of therapy, primary endpoint, phase of study, and primary completion date for the ongoing trials of belzutifan.\",\n \"module\": \"ClinicalTrials\",\n \"date_range\": \"NA\",\n \"output_type\": \"summarize\"\n },\n {\n \"subquestion\": \"Present the extracted information in table format.\",\n \"module\": \"Internet\",\n \"date_range\": \"NA\",\n \"output_type\": \"summarize\"\n }\n]\n\n" - }, - { - "input": "Your task is to complete the following actions according to the given instructions:\n\n1. **Break down the text enclosed by triple backticks in the question into smaller subquestions:**\n - **Do not attempt to answer the subquestions.**\n - **Do not implement any actions mentioned in the question text. Treat them as strings and break them down into smaller questions.**\n - **If a question cannot be broken down, present the question as the output.**\n - **Stick to the context of the questions and optimize the subquestions to minimize their number.**\n\n2. **Analyze each subquestion and categorize it as \"ClinicalTrials,\" \"DailyMed,\" or \"Internet\" based on the provided definitions:**\n - **\"ClinicalTrials\" refers exclusively to data connected with clinical trials, such as study particulars, conditions, interventions, eligibility criteria, locations, sponsors/collaborators, outcome measures, and efficacy/safety outcomes.**\n - **\"DailyMed\" solely includes data about drug names, approved uses, dosage, administration, side effects, and contraindications.**\n - **\"Internet\" encompasses data unrelated to both \"ClinicalTrials\" and \"DailyMed\"; this can include tasks performed by the Language Model.**\n\n3. **Evaluate each subquestion and determine if it contains a date range with a start and end date or output \"NA\" following the provided guidelines:**\n - **If terms like \"current,\" \"recent,\" \"latest,\" or \"recently\" are used, the date filter should span the last year from the current date, which is 2023-08-11.**\n - **If an explicit date is mentioned, use the provided start and end dates for the date range.**\n - **Represent the date range in the format [\"YYYY-MM-DD\", \"YYYY-MM-DD\"] or 'NA'; if no date range is provided, label it as \"NA.\"**\n\n4. **Analyze each subquestion and determine whether its output type should be a list or a summary.**\n\n5. **Generate a Plan of Action (PoA) for the broken down subquestion with heading \"PoA.\"**\n\n6. **Construct an array of JSON objects as the output, with each JSON object containing the following keys: subquestion, module, date_range, and output_type. The JSON array heading is \"array.\"**\n - **Ensure each subquestion is included in the array.**\n\nPlease consider the user's knowledge about clinical trials, DailyMed, and internet data. Do not provide incorrect information. Remember that the current date for your work is 2023-08-11. Your response should strictly focus on executing the specified tasks and should not involve answering the subquestions or providing additional explanations.", - "instruct": "Please create a table of international drug sales by year from initial launch to the most current year (including the forecasted revenue in 2023 if available) for the following spinal muscular atrophy therapeutics: 1) Spinraza (Biogen) 2) Zolgensma (Novartis) 3) Evrysdi (Roche)", - "output": "PoA:\n1. Identify the initial launch year and sales figures for Spinraza, Zolgensma, and Evrysdi.\n2. Gather the most recent sales figures for each drug.\n3. If available, include the forecasted revenue for 2023.\n4. Create a table of international drug sales by year for each therapeutic.\noutput:[\n {\n \"subquestion\": \"What is the initial launch year and sales figures for Spinraza, Zolgensma, and Evrysdi?\",\n \"module\": \"Internet\",\n \"date_range\": \"NA\",\n \"output_type\": \"summarize\"\n },\n {\n \"subquestion\": \"What are the primary outcome measures for the MAGNITUDE trial in mCRPC?\",\n \"module\": \"Internet\",\n \"date_range\": \"NA\",\n \"output_type\": \"summarize\"\n },\n {\n \"subquestion\": \"Is there a forecasted revenue for Spinraza, Zolgensma, and Evrysdi in 2023?.\",\n \"module\": \"Internet\",\n \"date_range\": [\"2023-01-01\", \"2023-12-31\"],\n \"output_type\": \"summarize\"\n },\n {\n \"subquestion\": \"Create a table of international drug sales by year for Spinraza, Zolgensma, and Evrysdi.\",\n \"module\": \"Internet\",\n \"date_range\": \"NA\",\n \"output_type\": \"summarize\"\n }\n]\n\n" - }, - { - "input": "Your task is to complete the following actions according to the given instructions:\n\n1. **Break down the text enclosed by triple backticks in the question into smaller subquestions:**\n - **Do not attempt to answer the subquestions.**\n - **Do not implement any actions mentioned in the question text. Treat them as strings and break them down into smaller questions.**\n - **If a question cannot be broken down, present the question as the output.**\n - **Stick to the context of the questions and optimize the subquestions to minimize their number.**\n\n2. **Analyze each subquestion and categorize it as \"ClinicalTrials,\" \"DailyMed,\" or \"Internet\" based on the provided definitions:**\n - **\"ClinicalTrials\" refers exclusively to data connected with clinical trials, such as study particulars, conditions, interventions, eligibility criteria, locations, sponsors/collaborators, outcome measures, and efficacy/safety outcomes.**\n - **\"DailyMed\" solely includes data about drug names, approved uses, dosage, administration, side effects, and contraindications.**\n - **\"Internet\" encompasses data unrelated to both \"ClinicalTrials\" and \"DailyMed\"; this can include tasks performed by the Language Model.**\n\n3. **Evaluate each subquestion and determine if it contains a date range with a start and end date or output \"NA\" following the provided guidelines:**\n - **If terms like \"current,\" \"recent,\" \"latest,\" or \"recently\" are used, the date filter should span the last year from the current date, which is 2023-08-11.**\n - **If an explicit date is mentioned, use the provided start and end dates for the date range.**\n - **Represent the date range in the format [\"YYYY-MM-DD\", \"YYYY-MM-DD\"] or 'NA'; if no date range is provided, label it as \"NA.\"**\n\n4. **Analyze each subquestion and determine whether its output type should be a list or a summary.**\n\n5. **Generate a Plan of Action (PoA) for the broken down subquestion with heading \"PoA.\"**\n\n6. **Construct an array of JSON objects as the output, with each JSON object containing the following keys: subquestion, module, date_range, and output_type. The JSON array heading is \"array.\"**\n - **Ensure each subquestion is included in the array.**\n\nPlease consider the user's knowledge about clinical trials, DailyMed, and internet data. Do not provide incorrect information. Remember that the current date for your work is 2023-08-11. Your response should strictly focus on executing the specified tasks and should not involve answering the subquestions or providing additional explanations.", - "instruct": "Compare PROpel and MAGNITUDE trial for mCRPC in terms of primary outcome measures in a tabular format", - "output": "PoA:\n1. Identify primary outcome measures for the PROpel trial in mCRPC.\n2. Identify primary outcome measures for the MAGNITUDE trial in mCRPC.\n3. Compare the primary outcome measures of the PROpel and MAGNITUDE trials in a tabular format.\noutput:[\n {\n \"subquestion\": \"What are the primary outcome measures for the PROpel trial in mCRPC?\",\n \"module\": \"ClinicalTrials\",\n \"date_range\": \"NA\",\n \"output_type\": \"List\"\n },\n {\n \"subquestion\": \"What are the primary outcome measures for the MAGNITUDE trial in mCRPC?\",\n \"module\": \"ClinicalTrials\",\n \"date_range\": \"NA\",\n \"output_type\": \"List\"\n },\n {\n \"subquestion\": \"Compare the primary outcome measures of the PROpel and MAGNITUDE trials in a tabular format.\",\n \"module\": \"Internet\",\n \"date_range\": \"NA\",\n \"output_type\": \"summarize\"\n }\n\n]\n\n" - }, - { - "input": "Your task is to complete the following actions according to the given instructions:\n\n1. **Break down the text enclosed by triple backticks in the question into smaller subquestions:**\n - **Do not attempt to answer the subquestions.**\n - **Do not implement any actions mentioned in the question text. Treat them as strings and break them down into smaller questions.**\n - **If a question cannot be broken down, present the question as the output.**\n - **Stick to the context of the questions and optimize the subquestions to minimize their number.**\n\n2. **Analyze each subquestion and categorize it as \"ClinicalTrials,\" \"DailyMed,\" or \"Internet\" based on the provided definitions:**\n - **\"ClinicalTrials\" refers exclusively to data connected with clinical trials, such as study particulars, conditions, interventions, eligibility criteria, locations, sponsors/collaborators, outcome measures, and efficacy/safety outcomes.**\n - **\"DailyMed\" solely includes data about drug names, approved uses, dosage, administration, side effects, and contraindications.**\n - **\"Internet\" encompasses data unrelated to both \"ClinicalTrials\" and \"DailyMed\"; this can include tasks performed by the Language Model.**\n\n3. **Evaluate each subquestion and determine if it contains a date range with a start and end date or output \"NA\" following the provided guidelines:**\n - **If terms like \"current,\" \"recent,\" \"latest,\" or \"recently\" are used, the date filter should span the last year from the current date, which is 2023-08-11.**\n - **If an explicit date is mentioned, use the provided start and end dates for the date range.**\n - **Represent the date range in the format [\"YYYY-MM-DD\", \"YYYY-MM-DD\"] or 'NA'; if no date range is provided, label it as \"NA.\"**\n\n4. **Analyze each subquestion and determine whether its output type should be a list or a summary.**\n\n5. **Generate a Plan of Action (PoA) for the broken down subquestion with heading \"PoA.\"**\n\n6. **Construct an array of JSON objects as the output, with each JSON object containing the following keys: subquestion, module, date_range, and output_type. The JSON array heading is \"array.\"**\n - **Ensure each subquestion is included in the array.**\n\nPlease consider the user's knowledge about clinical trials, DailyMed, and internet data. Do not provide incorrect information. Remember that the current date for your work is 2023-08-11. Your response should strictly focus on executing the specified tasks and should not involve answering the subquestions or providing additional explanations.", - "instruct": "List out the indications and approval dates for atezolizumab in a tabular format", - "output": "PoA:\n1. Identify indications for atezolizumab.\n2. Gather approval dates for atezolizumab.\n3. Organize the information in a tabular format.\noutput: [\n {\n \"subquestion\": \"What are the indications for atezolizumab?\",\n \"module\": \"DailyMed\",\n \"date_range\": \"NA\",\n \"output_type\": \"summarize\"\n },\n {\n \"subquestion\": \"What are the approval dates for atezolizumab?\",\n \"module\": \"Internet\",\n \"date_range\": \"NA\",\n \"output_type\": \"summarize\"\n },\n {\n \"subquestion\": \"Organize the indications and approval dates for atezolizumab in a tabular format.\",\n \"module\": \"Internet\",\n \"date_range\": \"NA\",\n \"output_type\": \"summarize\"\n }\n\n]\n\n" - }, - { - "input": "Your task is to complete the following actions according to the given instructions:\n\n1. **Break down the text enclosed by triple backticks in the question into smaller subquestions:**\n - **Do not attempt to answer the subquestions.**\n - **Do not implement any actions mentioned in the question text. Treat them as strings and break them down into smaller questions.**\n - **If a question cannot be broken down, present the question as the output.**\n - **Stick to the context of the questions and optimize the subquestions to minimize their number.**\n\n2. **Analyze each subquestion and categorize it as \"ClinicalTrials,\" \"DailyMed,\" or \"Internet\" based on the provided definitions:**\n - **\"ClinicalTrials\" refers exclusively to data connected with clinical trials, such as study particulars, conditions, interventions, eligibility criteria, locations, sponsors/collaborators, outcome measures, and efficacy/safety outcomes.**\n - **\"DailyMed\" solely includes data about drug names, approved uses, dosage, administration, side effects, and contraindications.**\n - **\"Internet\" encompasses data unrelated to both \"ClinicalTrials\" and \"DailyMed\"; this can include tasks performed by the Language Model.**\n\n3. **Evaluate each subquestion and determine if it contains a date range with a start and end date or output \"NA\" following the provided guidelines:**\n - **If terms like \"current,\" \"recent,\" \"latest,\" or \"recently\" are used, the date filter should span the last year from the current date, which is 2023-08-11.**\n - **If an explicit date is mentioned, use the provided start and end dates for the date range.**\n - **Represent the date range in the format [\"YYYY-MM-DD\", \"YYYY-MM-DD\"] or 'NA'; if no date range is provided, label it as \"NA.\"**\n\n4. **Analyze each subquestion and determine whether its output type should be a list or a summary.**\n\n5. **Generate a Plan of Action (PoA) for the broken down subquestion with heading \"PoA.\"**\n\n6. **Construct an array of JSON objects as the output, with each JSON object containing the following keys: subquestion, module, date_range, and output_type. The JSON array heading is \"array.\"**\n - **Ensure each subquestion is included in the array.**\n\nPlease consider the user's knowledge about clinical trials, DailyMed, and internet data. Do not provide incorrect information. Remember that the current date for your work is 2023-08-11. Your response should strictly focus on executing the specified tasks and should not involve answering the subquestions or providing additional explanations.", - "instruct": "Explain the best way to patent an antibody drug conjugate", - "output": "PoA:\n1. Research the best practices for patenting an antibody drug conjugate.\noutput: [\n {\n \"subquestion\": \"What is the best way to patent an antibody drug conjugate?\",\n \"module\": \"Internet\",\n \"date_range\": \"NA\",\n \"output_type\": \"summarize\"\n }\n]\n\n" - }, - { - "input": "Your task is to complete the following actions according to the given instructions:\n\n1. **Break down the text enclosed by triple backticks in the question into smaller subquestions:**\n - **Do not attempt to answer the subquestions.**\n - **Do not implement any actions mentioned in the question text. Treat them as strings and break them down into smaller questions.**\n - **If a question cannot be broken down, present the question as the output.**\n - **Stick to the context of the questions and optimize the subquestions to minimize their number.**\n\n2. **Analyze each subquestion and categorize it as \"ClinicalTrials,\" \"DailyMed,\" or \"Internet\" based on the provided definitions:**\n - **\"ClinicalTrials\" refers exclusively to data connected with clinical trials, such as study particulars, conditions, interventions, eligibility criteria, locations, sponsors/collaborators, outcome measures, and efficacy/safety outcomes.**\n - **\"DailyMed\" solely includes data about drug names, approved uses, dosage, administration, side effects, and contraindications.**\n - **\"Internet\" encompasses data unrelated to both \"ClinicalTrials\" and \"DailyMed\"; this can include tasks performed by the Language Model.**\n\n3. **Evaluate each subquestion and determine if it contains a date range with a start and end date or output \"NA\" following the provided guidelines:**\n - **If terms like \"current,\" \"recent,\" \"latest,\" or \"recently\" are used, the date filter should span the last year from the current date, which is 2023-08-11.**\n - **If an explicit date is mentioned, use the provided start and end dates for the date range.**\n - **Represent the date range in the format [\"YYYY-MM-DD\", \"YYYY-MM-DD\"] or 'NA'; if no date range is provided, label it as \"NA.\"**\n\n4. **Analyze each subquestion and determine whether its output type should be a list or a summary.**\n\n5. **Generate a Plan of Action (PoA) for the broken down subquestion with heading \"PoA.\"**\n\n6. **Construct an array of JSON objects as the output, with each JSON object containing the following keys: subquestion, module, date_range, and output_type. The JSON array heading is \"array.\"**\n - **Ensure each subquestion is included in the array.**\n\nPlease consider the user's knowledge about clinical trials, DailyMed, and internet data. Do not provide incorrect information. Remember that the current date for your work is 2023-08-11. Your response should strictly focus on executing the specified tasks and should not involve answering the subquestions or providing additional explanations.", - "instruct": "What are the ongoing clinical trials for anti-coagulants", - "output": "PoA:\n1. Identify ongoing clinical trials for anti-coagulants.\noutput: [\n{\n\"subquestion\": \"What are the ongoing clinical trials for anti-coagulants?\",\n\"module\": \"ClinicalTrials\",\n\"date_range\": [\"NA\", \"NA\"],\n\"output_type\": \"List\"\n}\n]\n" - }, - { - "input": "Your task is to complete the following actions according to the given instructions:\n\n1. **Break down the text enclosed by triple backticks in the question into smaller subquestions:**\n - **Do not attempt to answer the subquestions.**\n - **Do not implement any actions mentioned in the question text. Treat them as strings and break them down into smaller questions.**\n - **If a question cannot be broken down, present the question as the output.**\n - **Stick to the context of the questions and optimize the subquestions to minimize their number.**\n\n2. **Analyze each subquestion and categorize it as \"ClinicalTrials,\" \"DailyMed,\" or \"Internet\" based on the provided definitions:**\n - **\"ClinicalTrials\" refers exclusively to data connected with clinical trials, such as study particulars, conditions, interventions, eligibility criteria, locations, sponsors/collaborators, outcome measures, and efficacy/safety outcomes.**\n - **\"DailyMed\" solely includes data about drug names, approved uses, dosage, administration, side effects, and contraindications.**\n - **\"Internet\" encompasses data unrelated to both \"ClinicalTrials\" and \"DailyMed\"; this can include tasks performed by the Language Model.**\n\n3. **Evaluate each subquestion and determine if it contains a date range with a start and end date or output \"NA\" following the provided guidelines:**\n - **If terms like \"current,\" \"recent,\" \"latest,\" or \"recently\" are used, the date filter should span the last year from the current date, which is 2023-08-11.**\n - **If an explicit date is mentioned, use the provided start and end dates for the date range.**\n - **Represent the date range in the format [\"YYYY-MM-DD\", \"YYYY-MM-DD\"] or 'NA'; if no date range is provided, label it as \"NA.\"**\n\n4. **Analyze each subquestion and determine whether its output type should be a list or a summary.**\n\n5. **Generate a Plan of Action (PoA) for the broken down subquestion with heading \"PoA.\"**\n\n6. **Construct an array of JSON objects as the output, with each JSON object containing the following keys: subquestion, module, date_range, and output_type. The JSON array heading is \"array.\"**\n - **Ensure each subquestion is included in the array.**\n\nPlease consider the user's knowledge about clinical trials, DailyMed, and internet data. Do not provide incorrect information. Remember that the current date for your work is 2023-08-11. Your response should strictly focus on executing the specified tasks and should not involve answering the subquestions or providing additional explanations.", - "instruct": "What Is GDUFA? How does it help the FDA?", - "output": "PoA:\n1. Define GDUFA and its purpose.\n2. Explain how GDUFA helps the FDA.\noutput: [\n {\n \"subquestion\": \"What Is GDUFA?\",\n \"module\": \"Internet\",\n \"date_range\": \"NA\",\n \"output_type\": \"summarize\"\n },\n {\n \"subquestion\": \"How does GDUFA help the FDA?\",\n \"module\": \"Internet\",\n \"date_range\": \"NA\",\n \"output_type\": \"summarize\"\n }\n]\n" - }, - { - "input": "Your task is to complete the following actions according to the given instructions:\n\n1. **Break down the text enclosed by triple backticks in the question into smaller subquestions:**\n - **Do not attempt to answer the subquestions.**\n - **Do not implement any actions mentioned in the question text. Treat them as strings and break them down into smaller questions.**\n - **If a question cannot be broken down, present the question as the output.**\n - **Stick to the context of the questions and optimize the subquestions to minimize their number.**\n\n2. **Analyze each subquestion and categorize it as \"ClinicalTrials,\" \"DailyMed,\" or \"Internet\" based on the provided definitions:**\n - **\"ClinicalTrials\" refers exclusively to data connected with clinical trials, such as study particulars, conditions, interventions, eligibility criteria, locations, sponsors/collaborators, outcome measures, and efficacy/safety outcomes.**\n - **\"DailyMed\" solely includes data about drug names, approved uses, dosage, administration, side effects, and contraindications.**\n - **\"Internet\" encompasses data unrelated to both \"ClinicalTrials\" and \"DailyMed\"; this can include tasks performed by the Language Model.**\n\n3. **Evaluate each subquestion and determine if it contains a date range with a start and end date or output \"NA\" following the provided guidelines:**\n - **If terms like \"current,\" \"recent,\" \"latest,\" or \"recently\" are used, the date filter should span the last year from the current date, which is 2023-08-11.**\n - **If an explicit date is mentioned, use the provided start and end dates for the date range.**\n - **Represent the date range in the format [\"YYYY-MM-DD\", \"YYYY-MM-DD\"] or 'NA'; if no date range is provided, label it as \"NA.\"**\n\n4. **Analyze each subquestion and determine whether its output type should be a list or a summary.**\n\n5. **Generate a Plan of Action (PoA) for the broken down subquestion with heading \"PoA.\"**\n\n6. **Construct an array of JSON objects as the output, with each JSON object containing the following keys: subquestion, module, date_range, and output_type. The JSON array heading is \"array.\"**\n - **Ensure each subquestion is included in the array.**\n\nPlease consider the user's knowledge about clinical trials, DailyMed, and internet data. Do not provide incorrect information. Remember that the current date for your work is 2023-08-11. Your response should strictly focus on executing the specified tasks and should not involve answering the subquestions or providing additional explanations.", - "instruct": "Which drugs are approved for NSCLC KRAS G12C? What are their most recent sales numbers?", - "output": "PoA:\n1. Identify drugs approved for NSCLC KRAS G12C.\n2. Gather the most recent sales numbers for these drugs.\noutput: [\n {\n \"subquestion\": \"Which drugs are approved for NSCLC KRAS G12C?\",\n \"module\": \"DailyMed\",\n \"date_range\": [\"2022-01-01\", \"2023-03-31\"],\n \"output_type\": \"List\"\n },\n {\n \"subquestion\": \"What are their most recent sales numbers?\",\n \"module\": \"Internet\",\n \"date_range\": [\"2022-04-01\", \"2023-04-01\"],\n \"output_type\": \"summarize\"\n }\n ]" - }, - { - "input": "Your task is to complete the following actions according to the given instructions:\n\n1. **Break down the text enclosed by triple backticks in the question into smaller subquestions:**\n - **Do not attempt to answer the subquestions.**\n - **Do not implement any actions mentioned in the question text. Treat them as strings and break them down into smaller questions.**\n - **If a question cannot be broken down, present the question as the output.**\n - **Stick to the context of the questions and optimize the subquestions to minimize their number.**\n\n2. **Analyze each subquestion and categorize it as \"ClinicalTrials,\" \"DailyMed,\" or \"Internet\" based on the provided definitions:**\n - **\"ClinicalTrials\" refers exclusively to data connected with clinical trials, such as study particulars, conditions, interventions, eligibility criteria, locations, sponsors/collaborators, outcome measures, and efficacy/safety outcomes.**\n - **\"DailyMed\" solely includes data about drug names, approved uses, dosage, administration, side effects, and contraindications.**\n - **\"Internet\" encompasses data unrelated to both \"ClinicalTrials\" and \"DailyMed\"; this can include tasks performed by the Language Model.**\n\n3. **Evaluate each subquestion and determine if it contains a date range with a start and end date or output \"NA\" following the provided guidelines:**\n - **If terms like \"current,\" \"recent,\" \"latest,\" or \"recently\" are used, the date filter should span the last year from the current date, which is 2023-08-11.**\n - **If an explicit date is mentioned, use the provided start and end dates for the date range.**\n - **Represent the date range in the format [\"YYYY-MM-DD\", \"YYYY-MM-DD\"] or 'NA'; if no date range is provided, label it as \"NA.\"**\n\n4. **Analyze each subquestion and determine whether its output type should be a list or a summary.**\n\n5. **Generate a Plan of Action (PoA) for the broken down subquestion with heading \"PoA.\"**\n\n6. **Construct an array of JSON objects as the output, with each JSON object containing the following keys: subquestion, module, date_range, and output_type. The JSON array heading is \"array.\"**\n - **Ensure each subquestion is included in the array.**\n\nPlease consider the user's knowledge about clinical trials, DailyMed, and internet data. Do not provide incorrect information. Remember that the current date for your work is 2023-08-11. Your response should strictly focus on executing the specified tasks and should not involve answering the subquestions or providing additional explanations.", - "instruct": "Describe the latest developments in Alzheimer's treatment", - "output": "PoA:\n1. Gather information on the latest developments in Alzheimer's treatment.\noutput: [\n {\n \"subquestion\": \"What are the latest developments in Alzheimer's treatment?\",\n \"module\": \"Internet\",\n \"date_range\": \"NA\",\n \"output_type\": \"summarize\"\n }\n ]\n" - }, - { - "input": "Your task is to complete the following actions according to the given instructions:\n\n1. **Break down the text enclosed by triple backticks in the question into smaller subquestions:**\n - **Do not attempt to answer the subquestions.**\n - **Do not implement any actions mentioned in the question text. Treat them as strings and break them down into smaller questions.**\n - **If a question cannot be broken down, present the question as the output.**\n - **Stick to the context of the questions and optimize the subquestions to minimize their number.**\n\n2. **Analyze each subquestion and categorize it as \"ClinicalTrials,\" \"DailyMed,\" or \"Internet\" based on the provided definitions:**\n - **\"ClinicalTrials\" refers exclusively to data connected with clinical trials, such as study particulars, conditions, interventions, eligibility criteria, locations, sponsors/collaborators, outcome measures, and efficacy/safety outcomes.**\n - **\"DailyMed\" solely includes data about drug names, approved uses, dosage, administration, side effects, and contraindications.**\n - **\"Internet\" encompasses data unrelated to both \"ClinicalTrials\" and \"DailyMed\"; this can include tasks performed by the Language Model.**\n\n3. **Evaluate each subquestion and determine if it contains a date range with a start and end date or output \"NA\" following the provided guidelines:**\n - **If terms like \"current,\" \"recent,\" \"latest,\" or \"recently\" are used, the date filter should span the last year from the current date, which is 2023-08-11.**\n - **If an explicit date is mentioned, use the provided start and end dates for the date range.**\n - **Represent the date range in the format [\"YYYY-MM-DD\", \"YYYY-MM-DD\"] or 'NA'; if no date range is provided, label it as \"NA.\"**\n\n4. **Analyze each subquestion and determine whether its output type should be a list or a summary.**\n\n5. **Generate a Plan of Action (PoA) for the broken down subquestion with heading \"PoA.\"**\n\n6. **Construct an array of JSON objects as the output, with each JSON object containing the following keys: subquestion, module, date_range, and output_type. The JSON array heading is \"array.\"**\n - **Ensure each subquestion is included in the array.**\n\nPlease consider the user's knowledge about clinical trials, DailyMed, and internet data. Do not provide incorrect information. Remember that the current date for your work is 2023-08-11. Your response should strictly focus on executing the specified tasks and should not involve answering the subquestions or providing additional explanations.", - "instruct": "What are some recent deals in the psychedelic space?", - "output": "PoA:\n1. Identify recent deals in the psychedelic space.\noutput:[\n {\n \"subquestion\": \"What are some recent deals in the psychedelic space?\",\n \"module\": \"Internet\",\n \"date_range\": [\"2022-01-01\", \"now\"],\n \"output_type\": \"summarize\"\n }\n ]\n" - }, - { - "input": "Your task is to complete the following actions according to the given instructions:\n\n1. **Break down the text enclosed by triple backticks in the question into smaller subquestions:**\n - **Do not attempt to answer the subquestions.**\n - **Do not implement any actions mentioned in the question text. Treat them as strings and break them down into smaller questions.**\n - **If a question cannot be broken down, present the question as the output.**\n - **Stick to the context of the questions and optimize the subquestions to minimize their number.**\n\n2. **Analyze each subquestion and categorize it as \"ClinicalTrials,\" \"DailyMed,\" or \"Internet\" based on the provided definitions:**\n - **\"ClinicalTrials\" refers exclusively to data connected with clinical trials, such as study particulars, conditions, interventions, eligibility criteria, locations, sponsors/collaborators, outcome measures, and efficacy/safety outcomes.**\n - **\"DailyMed\" solely includes data about drug names, approved uses, dosage, administration, side effects, and contraindications.**\n - **\"Internet\" encompasses data unrelated to both \"ClinicalTrials\" and \"DailyMed\"; this can include tasks performed by the Language Model.**\n\n3. **Evaluate each subquestion and determine if it contains a date range with a start and end date or output \"NA\" following the provided guidelines:**\n - **If terms like \"current,\" \"recent,\" \"latest,\" or \"recently\" are used, the date filter should span the last year from the current date, which is 2023-08-11.**\n - **If an explicit date is mentioned, use the provided start and end dates for the date range.**\n - **Represent the date range in the format [\"YYYY-MM-DD\", \"YYYY-MM-DD\"] or 'NA'; if no date range is provided, label it as \"NA.\"**\n\n4. **Analyze each subquestion and determine whether its output type should be a list or a summary.**\n\n5. **Generate a Plan of Action (PoA) for the broken down subquestion with heading \"PoA.\"**\n\n6. **Construct an array of JSON objects as the output, with each JSON object containing the following keys: subquestion, module, date_range, and output_type. The JSON array heading is \"array.\"**\n - **Ensure each subquestion is included in the array.**\n\nPlease consider the user's knowledge about clinical trials, DailyMed, and internet data. Do not provide incorrect information. Remember that the current date for your work is 2023-08-11. Your response should strictly focus on executing the specified tasks and should not involve answering the subquestions or providing additional explanations.", - "instruct": "What's the latest news on Wegovy?", - "output": "PoA:\n1. Gather the most recent updates or news on Wegovy.\noutput:[\n {\n \"subquestion\": \"What is the latest news on Wegovy?\",\n \"module\": \"Internet\",\n \"date_range\": \"NA\",\n \"output_type\": \"summarize\"\n }\n ]\n" - }, - { - "input": "Your task is to complete the following actions according to the given instructions:\n\n1. **Break down the text enclosed by triple backticks in the question into smaller subquestions:**\n - **Do not attempt to answer the subquestions.**\n - **Do not implement any actions mentioned in the question text. Treat them as strings and break them down into smaller questions.**\n - **If a question cannot be broken down, present the question as the output.**\n - **Stick to the context of the questions and optimize the subquestions to minimize their number.**\n\n2. **Analyze each subquestion and categorize it as \"ClinicalTrials,\" \"DailyMed,\" or \"Internet\" based on the provided definitions:**\n - **\"ClinicalTrials\" refers exclusively to data connected with clinical trials, such as study particulars, conditions, interventions, eligibility criteria, locations, sponsors/collaborators, outcome measures, and efficacy/safety outcomes.**\n - **\"DailyMed\" solely includes data about drug names, approved uses, dosage, administration, side effects, and contraindications.**\n - **\"Internet\" encompasses data unrelated to both \"ClinicalTrials\" and \"DailyMed\"; this can include tasks performed by the Language Model.**\n\n3. **Evaluate each subquestion and determine if it contains a date range with a start and end date or output \"NA\" following the provided guidelines:**\n - **If terms like \"current,\" \"recent,\" \"latest,\" or \"recently\" are used, the date filter should span the last year from the current date, which is 2023-08-11.**\n - **If an explicit date is mentioned, use the provided start and end dates for the date range.**\n - **Represent the date range in the format [\"YYYY-MM-DD\", \"YYYY-MM-DD\"] or 'NA'; if no date range is provided, label it as \"NA.\"**\n\n4. **Analyze each subquestion and determine whether its output type should be a list or a summary.**\n\n5. **Generate a Plan of Action (PoA) for the broken down subquestion with heading \"PoA.\"**\n\n6. **Construct an array of JSON objects as the output, with each JSON object containing the following keys: subquestion, module, date_range, and output_type. The JSON array heading is \"array.\"**\n - **Ensure each subquestion is included in the array.**\n\nPlease consider the user's knowledge about clinical trials, DailyMed, and internet data. Do not provide incorrect information. Remember that the current date for your work is 2023-08-11. Your response should strictly focus on executing the specified tasks and should not involve answering the subquestions or providing additional explanations.", - "instruct": "What is the growth of US market for medicines in 2022?", - "output": "PoA:\n1. Determine the growth of the US market for medicines in 2022.\noutput: [\n {\n \"subquestion\": \"What is the growth of US market for medicines in 2022?\",\n \"module\": \"Internet\",\n \"date_range\": [\"2022-01-01\", \"2022-12-31\"],\n \"output_type\": \"summarize\"\n }\n ]\n\n" - }, - { - "input": "Your task is to complete the following actions according to the given instructions:\n\n1. **Break down the text enclosed by triple backticks in the question into smaller subquestions:**\n - **Do not attempt to answer the subquestions.**\n - **Do not implement any actions mentioned in the question text. Treat them as strings and break them down into smaller questions.**\n - **If a question cannot be broken down, present the question as the output.**\n - **Stick to the context of the questions and optimize the subquestions to minimize their number.**\n\n2. **Analyze each subquestion and categorize it as \"ClinicalTrials,\" \"DailyMed,\" or \"Internet\" based on the provided definitions:**\n - **\"ClinicalTrials\" refers exclusively to data connected with clinical trials, such as study particulars, conditions, interventions, eligibility criteria, locations, sponsors/collaborators, outcome measures, and efficacy/safety outcomes.**\n - **\"DailyMed\" solely includes data about drug names, approved uses, dosage, administration, side effects, and contraindications.**\n - **\"Internet\" encompasses data unrelated to both \"ClinicalTrials\" and \"DailyMed\"; this can include tasks performed by the Language Model.**\n\n3. **Evaluate each subquestion and determine if it contains a date range with a start and end date or output \"NA\" following the provided guidelines:**\n - **If terms like \"current,\" \"recent,\" \"latest,\" or \"recently\" are used, the date filter should span the last year from the current date, which is 2023-08-11.**\n - **If an explicit date is mentioned, use the provided start and end dates for the date range.**\n - **Represent the date range in the format [\"YYYY-MM-DD\", \"YYYY-MM-DD\"] or 'NA'; if no date range is provided, label it as \"NA.\"**\n\n4. **Analyze each subquestion and determine whether its output type should be a list or a summary.**\n\n5. **Generate a Plan of Action (PoA) for the broken down subquestion with heading \"PoA.\"**\n\n6. **Construct an array of JSON objects as the output, with each JSON object containing the following keys: subquestion, module, date_range, and output_type. The JSON array heading is \"array.\"**\n - **Ensure each subquestion is included in the array.**\n\nPlease consider the user's knowledge about clinical trials, DailyMed, and internet data. Do not provide incorrect information. Remember that the current date for your work is 2023-08-11. Your response should strictly focus on executing the specified tasks and should not involve answering the subquestions or providing additional explanations.", - "instruct": "List compounds that have been shown to induce thymic stromal lymphopoiten production in vitro", - "output": "PoA:\n1. Identify compounds that have been shown to induce thymic stromal lymphopoiten production in vitro.\noutput:[\n {\n \"subquestion\": \"What compounds have been shown to induce thymic stromal lymphopoietin production in vitro?\",\n \"module\": \"Internet\",\n \"date_range\": \"NA\",\n \"output_type\": \"summarize\"\n }\n]\n" - }, - { - "input": "Your task is to complete the following actions according to the given instructions:\n\n1. **Break down the text enclosed by triple backticks in the question into smaller subquestions:**\n - **Do not attempt to answer the subquestions.**\n - **Do not implement any actions mentioned in the question text. Treat them as strings and break them down into smaller questions.**\n - **If a question cannot be broken down, present the question as the output.**\n - **Stick to the context of the questions and optimize the subquestions to minimize their number.**\n\n2. **Analyze each subquestion and categorize it as \"ClinicalTrials,\" \"DailyMed,\" or \"Internet\" based on the provided definitions:**\n - **\"ClinicalTrials\" refers exclusively to data connected with clinical trials, such as study particulars, conditions, interventions, eligibility criteria, locations, sponsors/collaborators, outcome measures, and efficacy/safety outcomes.**\n - **\"DailyMed\" solely includes data about drug names, approved uses, dosage, administration, side effects, and contraindications.**\n - **\"Internet\" encompasses data unrelated to both \"ClinicalTrials\" and \"DailyMed\"; this can include tasks performed by the Language Model.**\n\n3. **Evaluate each subquestion and determine if it contains a date range with a start and end date or output \"NA\" following the provided guidelines:**\n - **If terms like \"current,\" \"recent,\" \"latest,\" or \"recently\" are used, the date filter should span the last year from the current date, which is 2023-08-11.**\n - **If an explicit date is mentioned, use the provided start and end dates for the date range.**\n - **Represent the date range in the format [\"YYYY-MM-DD\", \"YYYY-MM-DD\"] or 'NA'; if no date range is provided, label it as \"NA.\"**\n\n4. **Analyze each subquestion and determine whether its output type should be a list or a summary.**\n\n5. **Generate a Plan of Action (PoA) for the broken down subquestion with heading \"PoA.\"**\n\n6. **Construct an array of JSON objects as the output, with each JSON object containing the following keys: subquestion, module, date_range, and output_type. The JSON array heading is \"array.\"**\n - **Ensure each subquestion is included in the array.**\n\nPlease consider the user's knowledge about clinical trials, DailyMed, and internet data. Do not provide incorrect information. Remember that the current date for your work is 2023-08-11. Your response should strictly focus on executing the specified tasks and should not involve answering the subquestions or providing additional explanations.", - "instruct": "What are the key news updates in the allogenic CAR-T space in the last two weeks?", - "output": "PoA:\n1. Gather key news updates in the allogenic CAR-T space from the last two weeks.\noutput:[\n {\n \"subquestion\": \"What are the key news updates in the allogenic CAR-T space in the last two weeks?\",\n \"module\": \"Internet\",\n \"date_range\": [\"2023-08-01\", \"2023-08-11\"],\n \"output_type\": \"summarize\"\n }\n ]" - }, - { - "input": "Your task is to complete the following actions according to the given instructions:\n\n1. **Break down the text enclosed by triple backticks in the question into smaller subquestions:**\n - **Do not attempt to answer the subquestions.**\n - **Do not implement any actions mentioned in the question text. Treat them as strings and break them down into smaller questions.**\n - **If a question cannot be broken down, present the question as the output.**\n - **Stick to the context of the questions and optimize the subquestions to minimize their number.**\n\n2. **Analyze each subquestion and categorize it as \"ClinicalTrials,\" \"DailyMed,\" or \"Internet\" based on the provided definitions:**\n - **\"ClinicalTrials\" refers exclusively to data connected with clinical trials, such as study particulars, conditions, interventions, eligibility criteria, locations, sponsors/collaborators, outcome measures, and efficacy/safety outcomes.**\n - **\"DailyMed\" solely includes data about drug names, approved uses, dosage, administration, side effects, and contraindications.**\n - **\"Internet\" encompasses data unrelated to both \"ClinicalTrials\" and \"DailyMed\"; this can include tasks performed by the Language Model.**\n\n3. **Evaluate each subquestion and determine if it contains a date range with a start and end date or output \"NA\" following the provided guidelines:**\n - **If terms like \"current,\" \"recent,\" \"latest,\" or \"recently\" are used, the date filter should span the last year from the current date, which is 2023-08-11.**\n - **If an explicit date is mentioned, use the provided start and end dates for the date range.**\n - **Represent the date range in the format [\"YYYY-MM-DD\", \"YYYY-MM-DD\"] or 'NA'; if no date range is provided, label it as \"NA.\"**\n\n4. **Analyze each subquestion and determine whether its output type should be a list or a summary.**\n\n5. **Generate a Plan of Action (PoA) for the broken down subquestion with heading \"PoA.\"**\n\n6. **Construct an array of JSON objects as the output, with each JSON object containing the following keys: subquestion, module, date_range, and output_type. The JSON array heading is \"array.\"**\n - **Ensure each subquestion is included in the array.**\n\nPlease consider the user's knowledge about clinical trials, DailyMed, and internet data. Do not provide incorrect information. Remember that the current date for your work is 2023-08-11. Your response should strictly focus on executing the specified tasks and should not involve answering the subquestions or providing additional explanations.", - "instruct": "What is ctDNA? Can it be used for tumor-specific treatment?", - "output": "PoA:\n1. Define ctDNA and its significance.\n2. Explore the potential of ctDNA for tumor-specific treatment.\noutput:[\n {\n \"subquestion\": \"What is ctDNA?\",\n \"module\": \"Internet\",\n \"date_range\": \"NA\",\n \"output_type\": \"summarize\"\n },\n {\n \"subquestion\": \"Can it be used for tumor-specific treatment?\",\n \"module\": \"Internet\",\n \"date_range\": \"NA\",\n \"output_type\": \"summarize\"\n }\n ]" - }, - { - "input": "Your task is to complete the following actions according to the given instructions:\n\n1. **Break down the text enclosed by triple backticks in the question into smaller subquestions:**\n - **Do not attempt to answer the subquestions.**\n - **Do not implement any actions mentioned in the question text. Treat them as strings and break them down into smaller questions.**\n - **If a question cannot be broken down, present the question as the output.**\n - **Stick to the context of the questions and optimize the subquestions to minimize their number.**\n\n2. **Analyze each subquestion and categorize it as \"ClinicalTrials,\" \"DailyMed,\" or \"Internet\" based on the provided definitions:**\n - **\"ClinicalTrials\" refers exclusively to data connected with clinical trials, such as study particulars, conditions, interventions, eligibility criteria, locations, sponsors/collaborators, outcome measures, and efficacy/safety outcomes.**\n - **\"DailyMed\" solely includes data about drug names, approved uses, dosage, administration, side effects, and contraindications.**\n - **\"Internet\" encompasses data unrelated to both \"ClinicalTrials\" and \"DailyMed\"; this can include tasks performed by the Language Model.**\n\n3. **Evaluate each subquestion and determine if it contains a date range with a start and end date or output \"NA\" following the provided guidelines:**\n - **If terms like \"current,\" \"recent,\" \"latest,\" or \"recently\" are used, the date filter should span the last year from the current date, which is 2023-08-11.**\n - **If an explicit date is mentioned, use the provided start and end dates for the date range.**\n - **Represent the date range in the format [\"YYYY-MM-DD\", \"YYYY-MM-DD\"] or 'NA'; if no date range is provided, label it as \"NA.\"**\n\n4. **Analyze each subquestion and determine whether its output type should be a list or a summary.**\n\n5. **Generate a Plan of Action (PoA) for the broken down subquestion with heading \"PoA.\"**\n\n6. **Construct an array of JSON objects as the output, with each JSON object containing the following keys: subquestion, module, date_range, and output_type. The JSON array heading is \"array.\"**\n - **Ensure each subquestion is included in the array.**\n\nPlease consider the user's knowledge about clinical trials, DailyMed, and internet data. Do not provide incorrect information. Remember that the current date for your work is 2023-08-11. Your response should strictly focus on executing the specified tasks and should not involve answering the subquestions or providing additional explanations.", - "instruct": "Provide a summary of the company Origimed.", - "output": "PoA:\n1. Gather information on the company Origimed.\noutput:[\n {\n \"subquestion\": \"What is Origimed?\",\n \"module\": \"Internet\",\n \"date_range\": \"NA\",\n \"output_type\": \"summarize\"\n },\n {\n \"subquestion\": \"What are the products and services offered by Origimed?\",\n \"module\": \"Internet\",\n \"date_range\": \"NA\",\n \"output_type\": \"List\"\n },\n {\n \"subquestion\": \"Who are the key people behind Origimed?\",\n \"module\": \"Internet\",\n \"date_range\": \"NA\",\n \"output_type\": \"List\"\n },\n {\n \"subquestion\": \"What is the current status of Origimed?\",\n \"module\": \"Internet\",\n \"date_range\": \"NA\",\n \"output_type\": \"summarize\"\n }\n ]" - }, - { - "input": "Your task is to complete the following actions according to the given instructions:\n\n1. **Break down the text enclosed by triple backticks in the question into smaller subquestions:**\n - **Do not attempt to answer the subquestions.**\n - **Do not implement any actions mentioned in the question text. Treat them as strings and break them down into smaller questions.**\n - **If a question cannot be broken down, present the question as the output.**\n - **Stick to the context of the questions and optimize the subquestions to minimize their number.**\n\n2. **Analyze each subquestion and categorize it as \"ClinicalTrials,\" \"DailyMed,\" or \"Internet\" based on the provided definitions:**\n - **\"ClinicalTrials\" refers exclusively to data connected with clinical trials, such as study particulars, conditions, interventions, eligibility criteria, locations, sponsors/collaborators, outcome measures, and efficacy/safety outcomes.**\n - **\"DailyMed\" solely includes data about drug names, approved uses, dosage, administration, side effects, and contraindications.**\n - **\"Internet\" encompasses data unrelated to both \"ClinicalTrials\" and \"DailyMed\"; this can include tasks performed by the Language Model.**\n\n3. **Evaluate each subquestion and determine if it contains a date range with a start and end date or output \"NA\" following the provided guidelines:**\n - **If terms like \"current,\" \"recent,\" \"latest,\" or \"recently\" are used, the date filter should span the last year from the current date, which is 2023-08-11.**\n - **If an explicit date is mentioned, use the provided start and end dates for the date range.**\n - **Represent the date range in the format [\"YYYY-MM-DD\", \"YYYY-MM-DD\"] or 'NA'; if no date range is provided, label it as \"NA.\"**\n\n4. **Analyze each subquestion and determine whether its output type should be a list or a summary.**\n\n5. **Generate a Plan of Action (PoA) for the broken down subquestion with heading \"PoA.\"**\n\n6. **Construct an array of JSON objects as the output, with each JSON object containing the following keys: subquestion, module, date_range, and output_type. The JSON array heading is \"array.\"**\n - **Ensure each subquestion is included in the array.**\n\nPlease consider the user's knowledge about clinical trials, DailyMed, and internet data. Do not provide incorrect information. Remember that the current date for your work is 2023-08-11. Your response should strictly focus on executing the specified tasks and should not involve answering the subquestions or providing additional explanations.", - "instruct": "What papers are available on NRG1 fusions in cancer?", - "output": "PoA:\n1. Identify papers on NRG1 fusions in cancer.\noutput:[\n {\n \"subquestion\": \"What papers are available on NRG1 fusions in cancer?\",\n \"module\": \"Internet\",\n \"date_range\": \"NA\",\n \"output_type\": \"summarize\"\n }\n ]\n" - }, - { - "input": "Your task is to complete the following actions according to the given instructions:\n\n1. **Break down the text enclosed by triple backticks in the question into smaller subquestions:**\n - **Do not attempt to answer the subquestions.**\n - **Do not implement any actions mentioned in the question text. Treat them as strings and break them down into smaller questions.**\n - **If a question cannot be broken down, present the question as the output.**\n - **Stick to the context of the questions and optimize the subquestions to minimize their number.**\n\n2. **Analyze each subquestion and categorize it as \"ClinicalTrials,\" \"DailyMed,\" or \"Internet\" based on the provided definitions:**\n - **\"ClinicalTrials\" refers exclusively to data connected with clinical trials, such as study particulars, conditions, interventions, eligibility criteria, locations, sponsors/collaborators, outcome measures, and efficacy/safety outcomes.**\n - **\"DailyMed\" solely includes data about drug names, approved uses, dosage, administration, side effects, and contraindications.**\n - **\"Internet\" encompasses data unrelated to both \"ClinicalTrials\" and \"DailyMed\"; this can include tasks performed by the Language Model.**\n\n3. **Evaluate each subquestion and determine if it contains a date range with a start and end date or output \"NA\" following the provided guidelines:**\n - **If terms like \"current,\" \"recent,\" \"latest,\" or \"recently\" are used, the date filter should span the last year from the current date, which is 2023-08-11.**\n - **If an explicit date is mentioned, use the provided start and end dates for the date range.**\n - **Represent the date range in the format [\"YYYY-MM-DD\", \"YYYY-MM-DD\"] or 'NA'; if no date range is provided, label it as \"NA.\"**\n\n4. **Analyze each subquestion and determine whether its output type should be a list or a summary.**\n\n5. **Generate a Plan of Action (PoA) for the broken down subquestion with heading \"PoA.\"**\n\n6. **Construct an array of JSON objects as the output, with each JSON object containing the following keys: subquestion, module, date_range, and output_type. The JSON array heading is \"array.\"**\n - **Ensure each subquestion is included in the array.**\n\nPlease consider the user's knowledge about clinical trials, DailyMed, and internet data. Do not provide incorrect information. Remember that the current date for your work is 2023-08-11. Your response should strictly focus on executing the specified tasks and should not involve answering the subquestions or providing additional explanations.", - "instruct": "What is the preclinical pipeline for Harbour Biomed?", - "output": "PoA:\n1. Identify the preclinical pipeline for Harbour Biomed.\noutput:[\n {\n \"subquestion\": \"What is the preclinical pipeline for Harbour Biomed?\",\n \"module\": \"Internet\",\n \"date_range\": \"NA\",\n \"output_type\": \"summarize\"\n }\n ]\n" - }, - { - "input": "Your task is to complete the following actions according to the given instructions:\n\n1. **Break down the text enclosed by triple backticks in the question into smaller subquestions:**\n - **Do not attempt to answer the subquestions.**\n - **Do not implement any actions mentioned in the question text. Treat them as strings and break them down into smaller questions.**\n - **If a question cannot be broken down, present the question as the output.**\n - **Stick to the context of the questions and optimize the subquestions to minimize their number.**\n\n2. **Analyze each subquestion and categorize it as \"ClinicalTrials,\" \"DailyMed,\" or \"Internet\" based on the provided definitions:**\n - **\"ClinicalTrials\" refers exclusively to data connected with clinical trials, such as study particulars, conditions, interventions, eligibility criteria, locations, sponsors/collaborators, outcome measures, and efficacy/safety outcomes.**\n - **\"DailyMed\" solely includes data about drug names, approved uses, dosage, administration, side effects, and contraindications.**\n - **\"Internet\" encompasses data unrelated to both \"ClinicalTrials\" and \"DailyMed\"; this can include tasks performed by the Language Model.**\n\n3. **Evaluate each subquestion and determine if it contains a date range with a start and end date or output \"NA\" following the provided guidelines:**\n - **If terms like \"current,\" \"recent,\" \"latest,\" or \"recently\" are used, the date filter should span the last year from the current date, which is 2023-08-11.**\n - **If an explicit date is mentioned, use the provided start and end dates for the date range.**\n - **Represent the date range in the format [\"YYYY-MM-DD\", \"YYYY-MM-DD\"] or 'NA'; if no date range is provided, label it as \"NA.\"**\n\n4. **Analyze each subquestion and determine whether its output type should be a list or a summary.**\n\n5. **Generate a Plan of Action (PoA) for the broken down subquestion with heading \"PoA.\"**\n\n6. **Construct an array of JSON objects as the output, with each JSON object containing the following keys: subquestion, module, date_range, and output_type. The JSON array heading is \"array.\"**\n - **Ensure each subquestion is included in the array.**\n\nPlease consider the user's knowledge about clinical trials, DailyMed, and internet data. Do not provide incorrect information. Remember that the current date for your work is 2023-08-11. Your response should strictly focus on executing the specified tasks and should not involve answering the subquestions or providing additional explanations.", - "instruct": "Summarize the latest news in prostate cancer according to FiercePharma", - "output": "PoA:\n1. Gather the latest news in prostate cancer according to FiercePharma.\noutput:[\n {\n \"subquestion\": \"What is the latest news in prostate cancer according to FiercePharma?\",\n \"module\": \"Internet\",\n \"date_range\": \"NA\",\n \"output_type\": \"summarize\"\n }\n ]\n" - }, - { - "input": "Your task is to complete the following actions according to the given instructions:\n\n1. **Break down the text enclosed by triple backticks in the question into smaller subquestions:**\n - **Do not attempt to answer the subquestions.**\n - **Do not implement any actions mentioned in the question text. Treat them as strings and break them down into smaller questions.**\n - **If a question cannot be broken down, present the question as the output.**\n - **Stick to the context of the questions and optimize the subquestions to minimize their number.**\n\n2. **Analyze each subquestion and categorize it as \"ClinicalTrials,\" \"DailyMed,\" or \"Internet\" based on the provided definitions:**\n - **\"ClinicalTrials\" refers exclusively to data connected with clinical trials, such as study particulars, conditions, interventions, eligibility criteria, locations, sponsors/collaborators, outcome measures, and efficacy/safety outcomes.**\n - **\"DailyMed\" solely includes data about drug names, approved uses, dosage, administration, side effects, and contraindications.**\n - **\"Internet\" encompasses data unrelated to both \"ClinicalTrials\" and \"DailyMed\"; this can include tasks performed by the Language Model.**\n\n3. **Evaluate each subquestion and determine if it contains a date range with a start and end date or output \"NA\" following the provided guidelines:**\n - **If terms like \"current,\" \"recent,\" \"latest,\" or \"recently\" are used, the date filter should span the last year from the current date, which is 2023-08-11.**\n - **If an explicit date is mentioned, use the provided start and end dates for the date range.**\n - **Represent the date range in the format [\"YYYY-MM-DD\", \"YYYY-MM-DD\"] or 'NA'; if no date range is provided, label it as \"NA.\"**\n\n4. **Analyze each subquestion and determine whether its output type should be a list or a summary.**\n\n5. **Generate a Plan of Action (PoA) for the broken down subquestion with heading \"PoA.\"**\n\n6. **Construct an array of JSON objects as the output, with each JSON object containing the following keys: subquestion, module, date_range, and output_type. The JSON array heading is \"array.\"**\n - **Ensure each subquestion is included in the array.**\n\nPlease consider the user's knowledge about clinical trials, DailyMed, and internet data. Do not provide incorrect information. Remember that the current date for your work is 2023-08-11. Your response should strictly focus on executing the specified tasks and should not involve answering the subquestions or providing additional explanations.", - "instruct": "What are the most recent clinical trials to start recruiting for Keytruda?", - "output": "PoA:\n1. Identify the most recent clinical trials recruiting for Keytruda.\noutput:[\n {\n \"subquestion\": \"What are the most recent clinical trials to start recruiting for Keytruda?\",\n \"module\": \"ClinicalTrials\",\n \"date_range\": [\"2023-08-01\", \"2023-08-11\"],\n \"output_type\": \"List\"\n }\n ]\n\n" - }, - { - "input": "Your task is to complete the following actions according to the given instructions:\n\n1. **Break down the text enclosed by triple backticks in the question into smaller subquestions:**\n - **Do not attempt to answer the subquestions.**\n - **Do not implement any actions mentioned in the question text. Treat them as strings and break them down into smaller questions.**\n - **If a question cannot be broken down, present the question as the output.**\n - **Stick to the context of the questions and optimize the subquestions to minimize their number.**\n\n2. **Analyze each subquestion and categorize it as \"ClinicalTrials,\" \"DailyMed,\" or \"Internet\" based on the provided definitions:**\n - **\"ClinicalTrials\" refers exclusively to data connected with clinical trials, such as study particulars, conditions, interventions, eligibility criteria, locations, sponsors/collaborators, outcome measures, and efficacy/safety outcomes.**\n - **\"DailyMed\" solely includes data about drug names, approved uses, dosage, administration, side effects, and contraindications.**\n - **\"Internet\" encompasses data unrelated to both \"ClinicalTrials\" and \"DailyMed\"; this can include tasks performed by the Language Model.**\n\n3. **Evaluate each subquestion and determine if it contains a date range with a start and end date or output \"NA\" following the provided guidelines:**\n - **If terms like \"current,\" \"recent,\" \"latest,\" or \"recently\" are used, the date filter should span the last year from the current date, which is 2023-08-11.**\n - **If an explicit date is mentioned, use the provided start and end dates for the date range.**\n - **Represent the date range in the format [\"YYYY-MM-DD\", \"YYYY-MM-DD\"] or 'NA'; if no date range is provided, label it as \"NA.\"**\n\n4. **Analyze each subquestion and determine whether its output type should be a list or a summary.**\n\n5. **Generate a Plan of Action (PoA) for the broken down subquestion with heading \"PoA.\"**\n\n6. **Construct an array of JSON objects as the output, with each JSON object containing the following keys: subquestion, module, date_range, and output_type. The JSON array heading is \"array.\"**\n - **Ensure each subquestion is included in the array.**\n\nPlease consider the user's knowledge about clinical trials, DailyMed, and internet data. Do not provide incorrect information. Remember that the current date for your work is 2023-08-11. Your response should strictly focus on executing the specified tasks and should not involve answering the subquestions or providing additional explanations.", - "instruct": "What is MSI, and what is its significance in colorectal cancer?", - "output": "PoA:\n1. Define MSI and its significance in colorectal cancer.\noutput:[\n {\n \"subquestion\": \"What is MSI?\",\n \"module\": \"Internet\",\n \"date_range\": \"NA\",\n \"output_type\": \"summarize\"\n },\n {\n \"subquestion\": \"What is the significance of MSI in colorectal cancer?\",\n \"module\": \"Internet\",\n \"date_range\": \"NA\",\n \"output_type\": \"summarize\"\n }\n ]\n\n" - }, - { - "input": "Your task is to complete the following actions according to the given instructions:\n\n1. **Break down the text enclosed by triple backticks in the question into smaller subquestions:**\n - **Do not attempt to answer the subquestions.**\n - **Do not implement any actions mentioned in the question text. Treat them as strings and break them down into smaller questions.**\n - **If a question cannot be broken down, present the question as the output.**\n - **Stick to the context of the questions and optimize the subquestions to minimize their number.**\n\n2. **Analyze each subquestion and categorize it as \"ClinicalTrials,\" \"DailyMed,\" or \"Internet\" based on the provided definitions:**\n - **\"ClinicalTrials\" refers exclusively to data connected with clinical trials, such as study particulars, conditions, interventions, eligibility criteria, locations, sponsors/collaborators, outcome measures, and efficacy/safety outcomes.**\n - **\"DailyMed\" solely includes data about drug names, approved uses, dosage, administration, side effects, and contraindications.**\n - **\"Internet\" encompasses data unrelated to both \"ClinicalTrials\" and \"DailyMed\"; this can include tasks performed by the Language Model.**\n\n3. **Evaluate each subquestion and determine if it contains a date range with a start and end date or output \"NA\" following the provided guidelines:**\n - **If terms like \"current,\" \"recent,\" \"latest,\" or \"recently\" are used, the date filter should span the last year from the current date, which is 2023-08-11.**\n - **If an explicit date is mentioned, use the provided start and end dates for the date range.**\n - **Represent the date range in the format [\"YYYY-MM-DD\", \"YYYY-MM-DD\"] or 'NA'; if no date range is provided, label it as \"NA.\"**\n\n4. **Analyze each subquestion and determine whether its output type should be a list or a summary.**\n\n5. **Generate a Plan of Action (PoA) for the broken down subquestion with heading \"PoA.\"**\n\n6. **Construct an array of JSON objects as the output, with each JSON object containing the following keys: subquestion, module, date_range, and output_type. The JSON array heading is \"array.\"**\n - **Ensure each subquestion is included in the array.**\n\nPlease consider the user's knowledge about clinical trials, DailyMed, and internet data. Do not provide incorrect information. Remember that the current date for your work is 2023-08-11. Your response should strictly focus on executing the specified tasks and should not involve answering the subquestions or providing additional explanations.", - "instruct": "What is the size of the market for adult ITP treatment in the US?", - "output": "PoA:\n1. Determine the size of the market for adult ITP treatment in the US.\noutput:[\n {\n \"subquestion\": \"What is the size of the market for adult ITP treatment in the US?\",\n \"module\": \"Internet\",\n \"date_range\": \"NA\",\n \"output_type\": \"summarize\"\n }\n ]\n\n" - }, - { - "input": "Your task is to complete the following actions according to the given instructions:\n\n1. **Break down the text enclosed by triple backticks in the question into smaller subquestions:**\n - **Do not attempt to answer the subquestions.**\n - **Do not implement any actions mentioned in the question text. Treat them as strings and break them down into smaller questions.**\n - **If a question cannot be broken down, present the question as the output.**\n - **Stick to the context of the questions and optimize the subquestions to minimize their number.**\n\n2. **Analyze each subquestion and categorize it as \"ClinicalTrials,\" \"DailyMed,\" or \"Internet\" based on the provided definitions:**\n - **\"ClinicalTrials\" refers exclusively to data connected with clinical trials, such as study particulars, conditions, interventions, eligibility criteria, locations, sponsors/collaborators, outcome measures, and efficacy/safety outcomes.**\n - **\"DailyMed\" solely includes data about drug names, approved uses, dosage, administration, side effects, and contraindications.**\n - **\"Internet\" encompasses data unrelated to both \"ClinicalTrials\" and \"DailyMed\"; this can include tasks performed by the Language Model.**\n\n3. **Evaluate each subquestion and determine if it contains a date range with a start and end date or output \"NA\" following the provided guidelines:**\n - **If terms like \"current,\" \"recent,\" \"latest,\" or \"recently\" are used, the date filter should span the last year from the current date, which is 2023-08-11.**\n - **If an explicit date is mentioned, use the provided start and end dates for the date range.**\n - **Represent the date range in the format [\"YYYY-MM-DD\", \"YYYY-MM-DD\"] or 'NA'; if no date range is provided, label it as \"NA.\"**\n\n4. **Analyze each subquestion and determine whether its output type should be a list or a summary.**\n\n5. **Generate a Plan of Action (PoA) for the broken down subquestion with heading \"PoA.\"**\n\n6. **Construct an array of JSON objects as the output, with each JSON object containing the following keys: subquestion, module, date_range, and output_type. The JSON array heading is \"array.\"**\n - **Ensure each subquestion is included in the array.**\n\nPlease consider the user's knowledge about clinical trials, DailyMed, and internet data. Do not provide incorrect information. Remember that the current date for your work is 2023-08-11. Your response should strictly focus on executing the specified tasks and should not involve answering the subquestions or providing additional explanations.", - "instruct": "Show me all publications for mitapivat in the past 5 years.", - "output": "PoA:\n1. Identify publications for mitapivat in the past 5 years.\noutput:[\n {\n \"subquestion\": \"What are the publications related to mitapivat in the past 5 years?\",\n \"module\": \"Internet\",\n \"date_range\": [\"2018-01-01\", \"2023-08-11\"],\n \"output_type\": \"summarize\"\n }\n ]\n\n" - }, - { - "input": "Your task is to complete the following actions according to the given instructions:\n\n1. **Break down the text enclosed by triple backticks in the question into smaller subquestions:**\n - **Do not attempt to answer the subquestions.**\n - **Do not implement any actions mentioned in the question text. Treat them as strings and break them down into smaller questions.**\n - **If a question cannot be broken down, present the question as the output.**\n - **Stick to the context of the questions and optimize the subquestions to minimize their number.**\n\n2. **Analyze each subquestion and categorize it as \"ClinicalTrials,\" \"DailyMed,\" or \"Internet\" based on the provided definitions:**\n - **\"ClinicalTrials\" refers exclusively to data connected with clinical trials, such as study particulars, conditions, interventions, eligibility criteria, locations, sponsors/collaborators, outcome measures, and efficacy/safety outcomes.**\n - **\"DailyMed\" solely includes data about drug names, approved uses, dosage, administration, side effects, and contraindications.**\n - **\"Internet\" encompasses data unrelated to both \"ClinicalTrials\" and \"DailyMed\"; this can include tasks performed by the Language Model.**\n\n3. **Evaluate each subquestion and determine if it contains a date range with a start and end date or output \"NA\" following the provided guidelines:**\n - **If terms like \"current,\" \"recent,\" \"latest,\" or \"recently\" are used, the date filter should span the last year from the current date, which is 2023-08-11.**\n - **If an explicit date is mentioned, use the provided start and end dates for the date range.**\n - **Represent the date range in the format [\"YYYY-MM-DD\", \"YYYY-MM-DD\"] or 'NA'; if no date range is provided, label it as \"NA.\"**\n\n4. **Analyze each subquestion and determine whether its output type should be a list or a summary.**\n\n5. **Generate a Plan of Action (PoA) for the broken down subquestion with heading \"PoA.\"**\n\n6. **Construct an array of JSON objects as the output, with each JSON object containing the following keys: subquestion, module, date_range, and output_type. The JSON array heading is \"array.\"**\n - **Ensure each subquestion is included in the array.**\n\nPlease consider the user's knowledge about clinical trials, DailyMed, and internet data. Do not provide incorrect information. Remember that the current date for your work is 2023-08-11. Your response should strictly focus on executing the specified tasks and should not involve answering the subquestions or providing additional explanations.", - "instruct": "What are the top 5 drugs in rheumatoid arthritis by 2022 sales", - "output": "PoA:\n1. Identify drugs used for rheumatoid arthritis.\n2. Gather sales data for these drugs in 2022.\n3. Rank the drugs based on their sales figures.\n4. Select the top 5 drugs in rheumatoid arthritis by 2022 sales.\noutput:[\n {\n \"subquestion\": \"What are the top 5 drugs in rheumatoid arthritis by 2022 sales?\",\n \"module\": \"DailyMed\",\n \"date_range\": [\"2022-01-01\", \"2022-12-31\"],\n \"output_type\": \"List\"\n },\n {\n \"subquestion\": \"What are the sales figures for the top 5 drugs in rheumatoid arthritis by 2022 sales?\",\n \"module\": \"Internet\",\n \"date_range\": [\"2022-01-01\", \"2022-12-31\"],\n \"output_type\": \"summarize\"\n }\n]\n\n" - }, - { - "input": "Your task is to complete the following actions according to the given instructions:\n\n1. **Break down the text enclosed by triple backticks in the question into smaller subquestions:**\n - **Do not attempt to answer the subquestions.**\n - **Do not implement any actions mentioned in the question text. Treat them as strings and break them down into smaller questions.**\n - **If a question cannot be broken down, present the question as the output.**\n - **Stick to the context of the questions and optimize the subquestions to minimize their number.**\n\n2. **Analyze each subquestion and categorize it as \"ClinicalTrials,\" \"DailyMed,\" or \"Internet\" based on the provided definitions:**\n - **\"ClinicalTrials\" refers exclusively to data connected with clinical trials, such as study particulars, conditions, interventions, eligibility criteria, locations, sponsors/collaborators, outcome measures, and efficacy/safety outcomes.**\n - **\"DailyMed\" solely includes data about drug names, approved uses, dosage, administration, side effects, and contraindications.**\n - **\"Internet\" encompasses data unrelated to both \"ClinicalTrials\" and \"DailyMed\"; this can include tasks performed by the Language Model.**\n\n3. **Evaluate each subquestion and determine if it contains a date range with a start and end date or output \"NA\" following the provided guidelines:**\n - **If terms like \"current,\" \"recent,\" \"latest,\" or \"recently\" are used, the date filter should span the last year from the current date, which is 2023-08-11.**\n - **If an explicit date is mentioned, use the provided start and end dates for the date range.**\n - **Represent the date range in the format [\"YYYY-MM-DD\", \"YYYY-MM-DD\"] or 'NA'; if no date range is provided, label it as \"NA.\"**\n\n4. **Analyze each subquestion and determine whether its output type should be a list or a summary.**\n\n5. **Generate a Plan of Action (PoA) for the broken down subquestion with heading \"PoA.\"**\n\n6. **Construct an array of JSON objects as the output, with each JSON object containing the following keys: subquestion, module, date_range, and output_type. The JSON array heading is \"array.\"**\n - **Ensure each subquestion is included in the array.**\n\nPlease consider the user's knowledge about clinical trials, DailyMed, and internet data. Do not provide incorrect information. Remember that the current date for your work is 2023-08-11. Your response should strictly focus on executing the specified tasks and should not involve answering the subquestions or providing additional explanations.", - "instruct": "Which drugs does Dice Therapeutics have in clinical development?", - "output": "PoA:\n1. Identify drugs in clinical development by Dice Therapeutics.\noutput:[\n {\n \"subquestion\": \"What drugs does Dice Therapeutics have in clinical development?\",\n \"module\": \"ClinicalTrials\",\n \"date_range\": \"NA\",\n \"output_type\": \"List\"\n }\n ]" - }, - { - "input": "Your task is to complete the following actions according to the given instructions:\n\n1. **Break down the text enclosed by triple backticks in the question into smaller subquestions:**\n - **Do not attempt to answer the subquestions.**\n - **Do not implement any actions mentioned in the question text. Treat them as strings and break them down into smaller questions.**\n - **If a question cannot be broken down, present the question as the output.**\n - **Stick to the context of the questions and optimize the subquestions to minimize their number.**\n\n2. **Analyze each subquestion and categorize it as \"ClinicalTrials,\" \"DailyMed,\" or \"Internet\" based on the provided definitions:**\n - **\"ClinicalTrials\" refers exclusively to data connected with clinical trials, such as study particulars, conditions, interventions, eligibility criteria, locations, sponsors/collaborators, outcome measures, and efficacy/safety outcomes.**\n - **\"DailyMed\" solely includes data about drug names, approved uses, dosage, administration, side effects, and contraindications.**\n - **\"Internet\" encompasses data unrelated to both \"ClinicalTrials\" and \"DailyMed\"; this can include tasks performed by the Language Model.**\n\n3. **Evaluate each subquestion and determine if it contains a date range with a start and end date or output \"NA\" following the provided guidelines:**\n - **If terms like \"current,\" \"recent,\" \"latest,\" or \"recently\" are used, the date filter should span the last year from the current date, which is 2023-08-11.**\n - **If an explicit date is mentioned, use the provided start and end dates for the date range.**\n - **Represent the date range in the format [\"YYYY-MM-DD\", \"YYYY-MM-DD\"] or 'NA'; if no date range is provided, label it as \"NA.\"**\n\n4. **Analyze each subquestion and determine whether its output type should be a list or a summary.**\n\n5. **Generate a Plan of Action (PoA) for the broken down subquestion with heading \"PoA.\"**\n\n6. **Construct an array of JSON objects as the output, with each JSON object containing the following keys: subquestion, module, date_range, and output_type. The JSON array heading is \"array.\"**\n - **Ensure each subquestion is included in the array.**\n\nPlease consider the user's knowledge about clinical trials, DailyMed, and internet data. Do not provide incorrect information. Remember that the current date for your work is 2023-08-11. Your response should strictly focus on executing the specified tasks and should not involve answering the subquestions or providing additional explanations.", - "instruct": "Which phase of clinical development is DC-806 in?", - "output": "PoA:\n1. Identify the clinical development phase of DC-806.\noutput:[\n {\n \"subquestion\": \"What is the current phase of clinical development for DC-806?\",\n \"module\": \"ClinicalTrials\",\n \"date_range\": [\"2022-01-01\", \"now\"],\n \"output_type\": \"List\"\n }\n ]" - }, - { - "input": "Your task is to complete the following actions according to the given instructions:\n\n1. **Break down the text enclosed by triple backticks in the question into smaller subquestions:**\n - **Do not attempt to answer the subquestions.**\n - **Do not implement any actions mentioned in the question text. Treat them as strings and break them down into smaller questions.**\n - **If a question cannot be broken down, present the question as the output.**\n - **Stick to the context of the questions and optimize the subquestions to minimize their number.**\n\n2. **Analyze each subquestion and categorize it as \"ClinicalTrials,\" \"DailyMed,\" or \"Internet\" based on the provided definitions:**\n - **\"ClinicalTrials\" refers exclusively to data connected with clinical trials, such as study particulars, conditions, interventions, eligibility criteria, locations, sponsors/collaborators, outcome measures, and efficacy/safety outcomes.**\n - **\"DailyMed\" solely includes data about drug names, approved uses, dosage, administration, side effects, and contraindications.**\n - **\"Internet\" encompasses data unrelated to both \"ClinicalTrials\" and \"DailyMed\"; this can include tasks performed by the Language Model.**\n\n3. **Evaluate each subquestion and determine if it contains a date range with a start and end date or output \"NA\" following the provided guidelines:**\n - **If terms like \"current,\" \"recent,\" \"latest,\" or \"recently\" are used, the date filter should span the last year from the current date, which is 2023-08-11.**\n - **If an explicit date is mentioned, use the provided start and end dates for the date range.**\n - **Represent the date range in the format [\"YYYY-MM-DD\", \"YYYY-MM-DD\"] or 'NA'; if no date range is provided, label it as \"NA.\"**\n\n4. **Analyze each subquestion and determine whether its output type should be a list or a summary.**\n\n5. **Generate a Plan of Action (PoA) for the broken down subquestion with heading \"PoA.\"**\n\n6. **Construct an array of JSON objects as the output, with each JSON object containing the following keys: subquestion, module, date_range, and output_type. The JSON array heading is \"array.\"**\n - **Ensure each subquestion is included in the array.**\n\nPlease consider the user's knowledge about clinical trials, DailyMed, and internet data. Do not provide incorrect information. Remember that the current date for your work is 2023-08-11. Your response should strictly focus on executing the specified tasks and should not involve answering the subquestions or providing additional explanations.", - "instruct": "What are the strong forces present in high concentration protein formulations?", - "output": "PoA:\n1. Identify the strong forces present in high concentration protein formulations.\noutput:[\n {\n \"subquestion\": \"What are the strong forces present in high concentration protein formulations?\",\n \"module\": \"Internet\",\n \"date_range\": \"NA\",\n \"output_type\": \"summarize\"\n }\n ]\n\n" - }, - { - "input": "Your task is to complete the following actions according to the given instructions:\n\n1. **Break down the text enclosed by triple backticks in the question into smaller subquestions:**\n - **Do not attempt to answer the subquestions.**\n - **Do not implement any actions mentioned in the question text. Treat them as strings and break them down into smaller questions.**\n - **If a question cannot be broken down, present the question as the output.**\n - **Stick to the context of the questions and optimize the subquestions to minimize their number.**\n\n2. **Analyze each subquestion and categorize it as \"ClinicalTrials,\" \"DailyMed,\" or \"Internet\" based on the provided definitions:**\n - **\"ClinicalTrials\" refers exclusively to data connected with clinical trials, such as study particulars, conditions, interventions, eligibility criteria, locations, sponsors/collaborators, outcome measures, and efficacy/safety outcomes.**\n - **\"DailyMed\" solely includes data about drug names, approved uses, dosage, administration, side effects, and contraindications.**\n - **\"Internet\" encompasses data unrelated to both \"ClinicalTrials\" and \"DailyMed\"; this can include tasks performed by the Language Model.**\n\n3. **Evaluate each subquestion and determine if it contains a date range with a start and end date or output \"NA\" following the provided guidelines:**\n - **If terms like \"current,\" \"recent,\" \"latest,\" or \"recently\" are used, the date filter should span the last year from the current date, which is 2023-08-11.**\n - **If an explicit date is mentioned, use the provided start and end dates for the date range.**\n - **Represent the date range in the format [\"YYYY-MM-DD\", \"YYYY-MM-DD\"] or 'NA'; if no date range is provided, label it as \"NA.\"**\n\n4. **Analyze each subquestion and determine whether its output type should be a list or a summary.**\n\n5. **Generate a Plan of Action (PoA) for the broken down subquestion with heading \"PoA.\"**\n\n6. **Construct an array of JSON objects as the output, with each JSON object containing the following keys: subquestion, module, date_range, and output_type. The JSON array heading is \"array.\"**\n - **Ensure each subquestion is included in the array.**\n\nPlease consider the user's knowledge about clinical trials, DailyMed, and internet data. Do not provide incorrect information. Remember that the current date for your work is 2023-08-11. Your response should strictly focus on executing the specified tasks and should not involve answering the subquestions or providing additional explanations.", - "instruct": "What are the latest progresses on OncoTherad therapy?", - "output": "PoA:\n1. Gather the latest information on the progress of OncoTherad therapy.\noutput:[\n {\n \"subquestion\": \"What are the latest progresses on OncoTherad therapy?\",\n \"module\": \"Internet\",\n \"date_range\": \"NA\",\n \"output_type\": \"summarize\"\n }\n ]\n\n" - }, - { - "input": "Your task is to complete the following actions according to the given instructions:\n\n1. **Break down the text enclosed by triple backticks in the question into smaller subquestions:**\n - **Do not attempt to answer the subquestions.**\n - **Do not implement any actions mentioned in the question text. Treat them as strings and break them down into smaller questions.**\n - **If a question cannot be broken down, present the question as the output.**\n - **Stick to the context of the questions and optimize the subquestions to minimize their number.**\n\n2. **Analyze each subquestion and categorize it as \"ClinicalTrials,\" \"DailyMed,\" or \"Internet\" based on the provided definitions:**\n - **\"ClinicalTrials\" refers exclusively to data connected with clinical trials, such as study particulars, conditions, interventions, eligibility criteria, locations, sponsors/collaborators, outcome measures, and efficacy/safety outcomes.**\n - **\"DailyMed\" solely includes data about drug names, approved uses, dosage, administration, side effects, and contraindications.**\n - **\"Internet\" encompasses data unrelated to both \"ClinicalTrials\" and \"DailyMed\"; this can include tasks performed by the Language Model.**\n\n3. **Evaluate each subquestion and determine if it contains a date range with a start and end date or output \"NA\" following the provided guidelines:**\n - **If terms like \"current,\" \"recent,\" \"latest,\" or \"recently\" are used, the date filter should span the last year from the current date, which is 2023-08-11.**\n - **If an explicit date is mentioned, use the provided start and end dates for the date range.**\n - **Represent the date range in the format [\"YYYY-MM-DD\", \"YYYY-MM-DD\"] or 'NA'; if no date range is provided, label it as \"NA.\"**\n\n4. **Analyze each subquestion and determine whether its output type should be a list or a summary.**\n\n5. **Generate a Plan of Action (PoA) for the broken down subquestion with heading \"PoA.\"**\n\n6. **Construct an array of JSON objects as the output, with each JSON object containing the following keys: subquestion, module, date_range, and output_type. The JSON array heading is \"array.\"**\n - **Ensure each subquestion is included in the array.**\n\nPlease consider the user's knowledge about clinical trials, DailyMed, and internet data. Do not provide incorrect information. Remember that the current date for your work is 2023-08-11. Your response should strictly focus on executing the specified tasks and should not involve answering the subquestions or providing additional explanations.", - "instruct": "What is the recent news regarding Sanofi's development in multiple sclerosis?", - "output": "PoA:\n1. Gather the most recent news regarding Sanofi's development in multiple sclerosis.\noutput:[\n {\n \"subquestion\": \"What is the recent news regarding Sanofi's development in multiple sclerosis?\",\n \"module\": \"Internet\",\n \"date_range\": [\"2023-01-01\", \"2023-08-11\"],\n \"output_type\": \"summarize\"\n }\n ]\n\n" - } -] \ No newline at end of file