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

Knowledge Unlearning on WMDP Bio (test)

64.81Accuracy Forget

Refusal Training

23.043633.886844.7355.5732Jan 29, 2026
Updated 1mo ago

Evaluation Results

MethodLinks
2026.01
64.8160.39
2026.01
63.758.12
2026.01
63.1657.57
2026.01
62.7757.29
2026.01
62.6956.88
2026.01
31.8957.18
2026.01
29.460.63
2026.01
27.147.37
2026.01
25.8425.5
2026.01
25.6127.16
2026.01
24.6525.25