Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Question Selection on VQA-RAD (test)
Loading...
60.6
Risk
Oracle Bound
60.24
62.67
65.1
67.53
May 31, 2026
Risk
Accuracy
Rewrite
Updated 1d ago
Evaluation Results
Method
Method
Links
Risk
Accuracy
Rewrite
Oracle Bound
Strategy=Oracle Bound
2026.05
60.6
35.2
-
Risk-only Rerank
Strategy=Risk-only Rerank
2026.05
62.3
35.6
20.7
Risk + Margin
Strategy=Risk + Margin
2026.05
62.4
35.2
15.2
Utility Score
Strategy=Utility Score
2026.05
62.5
35.2
14.4
Risk + Intent/Length
Strategy=Risk + Intent...
2026.05
63.3
35.6
16.3
Original
Strategy=Original
2026.05
65.8
33.7
0
Prompt Rewrite
Strategy=Prompt Rewrite
2026.05
69.6
34.1
100
Feedback
Search any
task
Search any
task