yadu9989/memstrata-lme500-harness

Experimental, model-selectable LongMemEval-S verifier with immutable data, explicit scoring protocols and inspectable evidence.

2 best points · 0 comments · 1 Hacker News post · 0 GitHub stars · View on GitHub

Hacker News discussions

• How do you know your agent's memory layer is helping, not hurting? (2 points, 0 comments, by yadu989, 2026-10-16)