law-chat / README.md
leaderboard-pr-bot's picture
Adding Evaluation Results
a34e87f verified
|
raw
history blame
9.75 kB
metadata
language:
  - en
license: llama2
tags:
  - legal
datasets:
  - EleutherAI/pile
  - Open-Orca/OpenOrca
  - GAIR/lima
  - WizardLM/WizardLM_evol_instruct_V2_196k
metrics:
  - accuracy
pipeline_tag: text-generation
model-index:
  - name: law-chat
    results:
      - task:
          type: text-generation
          name: Text Generation
        dataset:
          name: AI2 Reasoning Challenge (25-Shot)
          type: ai2_arc
          config: ARC-Challenge
          split: test
          args:
            num_few_shot: 25
        metrics:
          - type: acc_norm
            value: 53.41
            name: normalized accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=AdaptLLM/law-chat
          name: Open LLM Leaderboard
      - task:
          type: text-generation
          name: Text Generation
        dataset:
          name: HellaSwag (10-Shot)
          type: hellaswag
          split: validation
          args:
            num_few_shot: 10
        metrics:
          - type: acc_norm
            value: 76.16
            name: normalized accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=AdaptLLM/law-chat
          name: Open LLM Leaderboard
      - task:
          type: text-generation
          name: Text Generation
        dataset:
          name: MMLU (5-Shot)
          type: cais/mmlu
          config: all
          split: test
          args:
            num_few_shot: 5
        metrics:
          - type: acc
            value: 50.24
            name: accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=AdaptLLM/law-chat
          name: Open LLM Leaderboard
      - task:
          type: text-generation
          name: Text Generation
        dataset:
          name: TruthfulQA (0-shot)
          type: truthful_qa
          config: multiple_choice
          split: validation
          args:
            num_few_shot: 0
        metrics:
          - type: mc2
            value: 43.53
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=AdaptLLM/law-chat
          name: Open LLM Leaderboard
      - task:
          type: text-generation
          name: Text Generation
        dataset:
          name: Winogrande (5-shot)
          type: winogrande
          config: winogrande_xl
          split: validation
          args:
            num_few_shot: 5
        metrics:
          - type: acc
            value: 75.45
            name: accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=AdaptLLM/law-chat
          name: Open LLM Leaderboard
      - task:
          type: text-generation
          name: Text Generation
        dataset:
          name: GSM8k (5-shot)
          type: gsm8k
          config: main
          split: test
          args:
            num_few_shot: 5
        metrics:
          - type: acc
            value: 18.5
            name: accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=AdaptLLM/law-chat
          name: Open LLM Leaderboard

Domain Adaptation of Large Language Models

This repo contains the domain-specific chat model developed from LLaMA-2-Chat-7B, using the method in our ICLR 2024 paper Adapting Large Language Models via Reading Comprehension.

We explore continued pre-training on domain-specific corpora for large language models. While this approach enriches LLMs with domain knowledge, it significantly hurts their prompting ability for question answering. Inspired by human learning via reading comprehension, we propose a simple method to transform large-scale pre-training corpora into reading comprehension texts, consistently improving prompting performance across tasks in biomedicine, finance, and law domains. Our 7B model competes with much larger domain-specific models like BloombergGPT-50B.

πŸ€— We are currently working hard on developing models across different domains, scales and architectures! Please stay tuned! πŸ€—

**************************** Updates ****************************

Domain-Specific LLaMA-1

LLaMA-1-7B

In our paper, we develop three domain-specific models from LLaMA-1-7B, which are also available in Huggingface: Biomedicine-LLM, Finance-LLM and Law-LLM, the performances of our AdaptLLM compared to other domain-specific LLMs are:

LLaMA-1-13B

Moreover, we scale up our base model to LLaMA-1-13B to see if our method is similarly effective for larger-scale models, and the results are consistently positive too: Biomedicine-LLM-13B, Finance-LLM-13B and Law-LLM-13B.

Domain-Specific LLaMA-2-Chat

Our method is also effective for aligned models! LLaMA-2-Chat requires a specific data format, and our reading comprehension can perfectly fit the data format by transforming the reading comprehension into a multi-turn conversation. We have also open-sourced chat models in different domains: Biomedicine-Chat, Finance-Chat and Law-Chat

For example, to chat with the law-chat model:

from transformers import AutoModelForCausalLM, AutoTokenizer

model = AutoModelForCausalLM.from_pretrained("AdaptLLM/law-chat")
tokenizer = AutoTokenizer.from_pretrained("AdaptLLM/law-chat")

# Put your input here:
user_input = '''Question: Which of the following is false about ex post facto laws?
Options:
- They make criminal an act that was innocent when committed.
- They prescribe greater punishment for an act than was prescribed when it was done.
- They increase the evidence required to convict a person than when the act was done.
- They alter criminal offenses or punishment in a substantially prejudicial manner for the purpose of punishing a person for some past activity.

Please provide your choice first and then provide explanations if possible.'''

# Apply the prompt template and system prompt of LLaMA-2-Chat demo for chat models (NOTE: NO prompt template is required for base models!)
our_system_prompt = "\nYou are a helpful, respectful and honest assistant. Always answer as helpfully as possible, while being safe.  Your answers should not include any harmful, unethical, racist, sexist, toxic, dangerous, or illegal content. Please ensure that your responses are socially unbiased and positive in nature.\n\nIf a question does not make any sense, or is not factually coherent, explain why instead of answering something not correct. If you don't know the answer to a question, please don't share false information.\n" # Please do NOT change this
prompt = f"<s>[INST] <<SYS>>{our_system_prompt}<</SYS>>\n\n{user_input} [/INST]"

# # NOTE:
# # If you want to apply your own system prompt, please integrate it into the instruction part following our system prompt like this:
# your_system_prompt = "Please, answer this question faithfully."
# prompt = f"<s>[INST] <<SYS>>{our_system_prompt}<</SYS>>\n\n{your_system_prompt}\n{user_input} [/INST]"

inputs = tokenizer(prompt, return_tensors="pt", add_special_tokens=False).input_ids.to(model.device)
outputs = model.generate(input_ids=inputs, max_length=4096)[0]

answer_start = int(inputs.shape[-1])
pred = tokenizer.decode(outputs[answer_start:], skip_special_tokens=True)

print(f'### User Input:\n{user_input}\n\n### Assistant Output:\n{pred}')

Domain-Specific Tasks

To easily reproduce our results, we have uploaded the filled-in zero/few-shot input instructions and output completions of each domain-specific task: biomedicine-tasks, finance-tasks, and law-tasks.

Note: those filled-in instructions are specifically tailored for models before alignment and do NOT fit for the specific data format required for chat models.

Citation

If you find our work helpful, please cite us:

@article{adaptllm,
  title={Adapting large language models via reading comprehension},
  author={Cheng, Daixuan and Huang, Shaohan and Wei, Furu},
  journal={arXiv preprint arXiv:2309.09530},
  year={2023}
}

Open LLM Leaderboard Evaluation Results

Detailed results can be found here

Metric Value
Avg. 52.88
AI2 Reasoning Challenge (25-Shot) 53.41
HellaSwag (10-Shot) 76.16
MMLU (5-Shot) 50.24
TruthfulQA (0-shot) 43.53
Winogrande (5-shot) 75.45
GSM8k (5-shot) 18.50