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

Over-Refusal Evaluation on OrBench-Hard-1k, FalseReject, MorBench, and OverBench

23.22ORR (OrBench-Hard-1k)

YuFeng-XGuard-0.6B

22.014430.152238.2946.4278Jun 11, 2026
Updated 1mo ago

Evaluation Results

MethodLinks
2026.06
23.2210.723.085.4610.62
2026.06
24.445.364.682.999.37
2026.06
24.445.634.792.349.3
2026.06
28.316.186.973.2611.27
2026.06
30.112.617.455.5816.43
2026.06
31.579.655.025.4612.93
2026.06
37.6810.1910.953.915.68
2026.06
43.587.5111.524.3816.75
2026.06
45.829.1213.688.8919.38
2026.06
48.729.3815.2810.120.87
2026.06
53.3610.4619.275.222.07