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

Evidence-page retrieval on MMLongBench-Doc

75.68Recall

EviProp

32.811243.940655.0766.1994Jun 8, 2026
Updated 1mo ago

Evaluation Results

MethodLinks
2026.06
75.6823.4768.0869.83
2026.06
71.3522.3464.3766.4
2026.06
68.7721.1560.4861.85
2026.06
68.5533.6565.3868.68
2026.06
64.1531.5861.4865.01
2026.06
61.9519.0453.9155.36
2026.06
60.2129.5557.160.4
2026.06
53.312650.3753.59
2026.06
45.4159.6759.6759.67
2026.06
43.2656.4456.4456.44
2026.06
39.3451.4151.4151.41
2026.06
34.4645.5545.5545.55