https://wandb.ai/open-assistant/supervised-finetuning/runs/i9gmn0dt
Trained with residual dropout 0.1
Open LLM Leaderboard Evaluation Results
Detailed results can be found here
Metric | Value |
---|---|
Avg. | 49.28 |
ARC (25-shot) | 56.4 |
HellaSwag (10-shot) | 79.34 |
MMLU (5-shot) | 46.59 |
TruthfulQA (0-shot) | 48.6 |
Winogrande (5-shot) | 75.22 |
GSM8K (5-shot) | 11.83 |
DROP (3-shot) | 27.03 |