Confabulations Leaderboard

Evals

Benchmark / dataset

RAG confabulation rates on misleading questions.

Category
Hallucination & factuality
Insurance exposure
Hallucination
Publisher
Independent (Lech Mazur)
Publisher type
Company (open source)
License
unknown
Status
Stale (>12mo)
Last commit
2025-08-07
Stars
249

View source on GitHub →

Wanich Risk Lab, 2026.