PersistBench

Evals

Benchmark / dataset

Memory risks: cross-domain leakage and sycophancy from stored memories.

Category
Agent safety & harmful actions
Insurance exposure
Privacy
Publisher
Contributed to inspect_evals
Publisher type
Academic
License
MIT
Status
In inspect_evals (active)
Last commit
2026-10-15
Stars
704(shared repo star count — UKGovernmentBEIS/inspect_evals backs 19 entries in this catalog)

View source on GitHub →


Wanich Risk Lab, 2026.