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

Fine-tuning on Average (GSM8k, ViGGO, SQL) (val)

81.1Accuracy

Full

44.28453.84263.472.958Jun 14, 2026
Updated 1mo ago

Evaluation Results

MethodLinks
2026.06
81.1
2026.06
77.6
2026.06
77.5
2026.06
76.5
2026.06
74.9
2026.06
74
2026.06
71
2026.06
70.4
2026.06
69.2
2026.06
65.8
2026.06
63.9
2026.06
54.9
2026.06
54.6
2026.06
52.3
2026.06
52
2026.06
45.7