What's up with the MATH Lvl 5 score on HF Open LLM Leaderboard 2?

#16
by invalid-access - opened

Llama-3.1-70B-Instruct only scores 2.72 on the MATH Lvl 5 metric on HF Open LLM Leaderboard 2. Something needs fixing?

Screenshot taken as of date:
Screenshot 2024-07-27 at 2.37.15 PM.png

invalid-access changed discussion title from What's up with the MATH Lvl5 score on HF Open LLM Leaderboard 2? to What's up with the MATH Lvl 5 score on HF Open LLM Leaderboard 2?

Very curious about this too. Surprised more people aren't questioning this. I wonder if this is a parsing error of the results and not the actual score.

Sign up or log in to comment