Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

General NLP Evaluation on The Pile Downstream Evaluation Suite

29.7HellaSwag Accuracy

DOGE

29.28429.39229.529.608Apr 9, 2026
Updated 9d ago

Evaluation Results

MethodLinks
2026.04
29.756.929.26425.550.661.740.611.941.1
2026.04
29.658.829.46625.951.16139.112.641.5
2026.04
29.658.429.86624.951.461.140.512.841.6
2026.04
29.660.5296729.751.465.240.215.143.1
2026.04
29.459.529.466.525.153.662.541.212.342.2
2026.04
29.358.129.666.12552.562.739.912.241.7
2026.04
29.358.430.265.925.651.361.340.212.141.6