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

Paper, table2 #13

Closed
serotoninpm opened this issue Jun 28, 2023 · 2 comments
Closed

Paper, table2 #13

serotoninpm opened this issue Jun 28, 2023 · 2 comments

Comments

@serotoninpm
Copy link

I am impressed with your research. Thank you for your good research.

But I have a question and would like to ask.

According to Table 2 of the paper, success and failure modes are divided.

  1. what is the definition of success mode and failure mode?
  2. if success mode is a successful case, it should not include false positives, because false positives are predicting something wrong as right.
  3. ultimately, Hallucinated reasoning traces or facts are present in both success mode and failure mode. I wonder why?
table2

Thanks!

@ysymyth
Copy link
Owner

ysymyth commented Jul 2, 2023

Thanks @serotoninpm !

  1. Success and failure are simply defined as whether the LLM is able to reach correct answers after the ReAct procedure (measured by the final response).
  2. Therefore, the LM can reach correct answers for the wrong ReAct trajectory. We would like to distinguish such cases from those successful cases where the trajectories are also correct, therefore report "false positive" in the table.
  3. Hallucination can happen anytime, both in success and failure cases (even the model hallucinates at some point, it may still be able to reach the correct answer for the wrong reason). Therefore in both modes we report the hallucination traces.

@ysymyth ysymyth closed this as completed Jul 4, 2023
@serotoninpm
Copy link
Author

Thank you for your detailed and kind explanation!

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