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

Intent Leakage on BioQA

24.25ASR@1

Soft-PO

10.594814.139917.68521.2301Apr 25, 2026
Updated 1mo ago

Evaluation Results

MethodLinks
2026.04
24.2542.8433.65
2026.04
21.6740.1331.26
2026.04
21.4839.6531.6
2026.04
20.5638.7230.77
2026.04
19.8937.9930.03
2026.04
19.3337.0229.21
2026.04
18.4136.1428.43
2026.04
14.1631.9825.76
2026.04
13.4431.0524.78
2026.04
12.7130.1424.16
2026.04
11.9929.3723.26
2026.04
11.1228.5722.52