dial481/locomo-audit

Full audit of the LoCoMo benchmark

3 best points · 3 comments · 1 Hacker News post · 23 GitHub stars · View on GitHub

Hacker News discussions

• LoCoMo AI Benchmark: 6.4% of answer key wrong, judge accepts 63% of fake answers (3 points, 3 comments, by dial481, 2026-03-23)