SycophancyEval

Evals

Benchmark / dataset

Sycophancy across free-form tasks.

Category
Deception, scheming & alignment auditing
Insurance exposure
Hallucination
Publisher
Anthropic (paper)
Publisher type
Frontier lab
License
unknown
Status
Stale (>12mo)
Last commit
2023-10-25
Stars
136

View source on GitHub →

Wanich Risk Lab, 2026.