StrongREJECT

Evals

Benchmark / dataset

Jailbreak benchmark with calibrated grader.

Category
Jailbreak & misuse
Insurance exposure
Regulatory/bias
Publisher
UC Berkeley
Publisher type
Academic
License
MIT
Status
Stale (>12mo)
Last commit
2025-07-07
Stars
160
Notes
Supersedes alexandrasouly/strongreject.

View source on GitHub →


Wanich Risk Lab, 2026.