You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Since GPT-4 and ChatGPT API replies may change every day, week or month, there may be no way to get the same output as reported 8 months ago (according to the last change date of the README_zh.md link you provided).
You can also try to find exact version of lm-eval code they used (probably some 0.3.0 like with updates) and check reproducibility.
As https://github.com/hkust-nlp/ceval/blob/main/README_zh.md shows:
but the result of ours is:
The script is:
How can I get the same ouput as the official version?
Thanks!
The text was updated successfully, but these errors were encountered: