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

Confusion about notation in Table 2 #8

Open
JINAILAB opened this issue Jan 7, 2025 · 1 comment
Open

Confusion about notation in Table 2 #8

JINAILAB opened this issue Jan 7, 2025 · 1 comment

Comments

@JINAILAB
Copy link

JINAILAB commented Jan 7, 2025

While reviewing Table 2, I noticed the notation for "SFT w/ Complex CoT (x̂, ŷ)". However, based on the context and the explanation in the paper, it seems that (x̂, ŷ) might actually refer to (ê, ŷ).

Could you please clarify if this is a typographical error or if I am misunderstanding the notation?

@jymChen
Copy link
Contributor

jymChen commented Jan 20, 2025

Hi @JINAILAB,

This is not a typographical error. "ê" represents the model’s internal reasoning, while "ŷ" is the complete answer based on that reasoning. The pair (ê, ŷ) is somewhat analogous to OpenAI-o1's hidden thought process (which you cannot access) and its final response. Therefore, even without ê, ŷ still constitutes a full answer. "SFT w/ Complex CoT (x̂, ŷ)" is more akin to directly distilling o1's response rather than learning its thought process.

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