Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Why isn't the <think> and <answer> tag used in the data's prompt? #19

Open
OKC13 opened this issue Feb 13, 2025 · 2 comments
Open

Why isn't the <think> and <answer> tag used in the data's prompt? #19

OKC13 opened this issue Feb 13, 2025 · 2 comments

Comments

@OKC13
Copy link

OKC13 commented Feb 13, 2025

As far as I know, the technical report of deepseek R1 suggests the usage as follows: "The reasoning process and answer are enclosed within and tags." May I know the reason for changing the prompt format? And does it have any impact on the results?

dataset from this project

Image

dataset from Logic-RL

Image

@OKC13
Copy link
Author

OKC13 commented Feb 13, 2025

Image

@michaelzhiluo
Copy link
Contributor

We simply follow Deepseek's repo's recommendation: https://github.com/deepseek-ai/DeepSeek-R1?tab=readme-ov-file#usage-recommendations

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants