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

Generative Task on PubMedQA labeled (0.8:0.2)

76.8Accuracy

LoRA + SAM

18.35233.52648.763.874Jun 22, 2026
Updated 1mo ago

Evaluation Results

MethodLinks
2026.06
76.8
2026.06
76.7
2026.06
76.58
2026.06
76.1
2026.06
75.95
2026.06
75.65
2026.06
75.2
2026.06
74.66
2026.06
74.58
2026.06
74.2
2026.06
74.15
2026.06
73.8
2026.06
73.65
2026.06
73.15
2026.06
72.6
2026.06
71.62
2026.06
70.95
2026.06
70.05
2026.06
69.9
2026.06
68.15
2026.06
67.9
2026.06
67.6
2026.06
67.55
2026.06
66.95
2026.06
66.4
2026.06
64.86
2026.06
46.05
2026.06
20.6