Open LLM Leaderboard results
#3
by
SaisExperiments
- opened
Why does this model score so different on the official leaderboard?
| Metric | Model Card Results | Leaderboard Results |
|---|---|---|
| Avg. | 29.79 | 15.4 |
| IFEval (0-Shot) | 32.12 | 31.54 |
| BBH (3-Shot) | 42.23 | 19.53 |
| MATH Lvl 5 (4-Shot) | 8.16 | 7.63 |
| GPQA (0-shot) | 27.10 | 3.69 |
| MuSR (0-shot) | 40.61 | 9.41 |
| MMLU-PRO (5-shot) | 28.49 | 20.6 |
Hello, this is a confusion that I made during the report (forgot to normalize the score, im adding the new score from the leaderboard rn)
qnguyen3
changed discussion status to
closed