Spaces:
Running
🚩 Report
fake
not a o1
It's GPT 4
Did you ask the model which version it is? The chatbot here doesn't use a system prompt, so it's not aware of its own version. For more details on this issue, you can refer to a response I gave to a previous chatbot we built: https://huggingface.co/spaces/yuntian-deng/ChatGPT4/discussions/1.
That said, I understand it can be difficult to trust the claimed API version when using a third-party chatbot like ours. If I were to test whether a chatbot is based on o1 or gpt-4o, I would use one of the following tests:
- Ask, "How many 'r's are there in the word 'strawberry'?"
- Ask, "Which is bigger, 9.11 or 9.9?"
Both tests were originally suggested by others, but I can't locate the sources at the moment. The second question was proposed by my colleague at AI2, Yuchen Lin.
Just ask what company is made by and what is their knowledge cutoff date