[KongchangAI]
BenchmarkWeapons of Mass Destruction Proxy

WMDP

A benchmark focused on unlearning dangerous knowledge (biological, chemical, cyberattack, and similar domains), measuring whether a model retains general capabilities after harmful capabilities are removed.

Timeline (last 90 days)

Nov 16

Experiments on the TOFU and WMDP benchmarks show that UUE and UUE-L significantly outperform existing methods in unlearning effectiveness.

Unverified50%
Nov 16

WMDP (Weapons of Mass Destruction Proxy) focuses on unlearning dangerous knowledge (biological, chemical, cyberattacks, etc.) and measures whether a model retains its general capabilities after removing harmful abilities.

Unverified50%

All Facts (2)

Source Articles