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

Zero-shot Classification on ARC-Challenge, ARC-Easy, PIQA, and WinoGrande

73.9Accuracy

Full-precision

37.18846.71956.2565.781Mar 30, 2026
Updated 18d ago

Evaluation Results

MethodLinks
2026.03
73.9
2026.03
73.9
2026.03
73.6
2026.03
72.3
2026.03
72
2026.03
71.5
2026.03
70.7
2026.03
70.5
2026.03
70.1
2026.03
69.8
2026.03
69.7
2026.03
69.7
2026.03
69.5
2026.03
67.6
2026.03
67.2
2026.03
66.8
2026.03
66
2026.03
65.4
2026.03
65.1
2026.03
65
2026.03
58.3
2026.03
58.1
2026.03
53.3
2026.03
52.1
2026.03
50.6
2026.03
49.6
2026.03
38.6