Show HN: LLM memory without context bleed; 100% precision vs. <10% vector searchtenureai.dev4 ポイント·投稿者 decorner·先月·1 コメント
Why Vector Search fails at LLM memory (and a benchmark to prove it)github.com3 ポイント·投稿者 decorner·先月·0 コメント
LLM memory systems benchmark: high recall near-zero precision for tested systemsarxiv.org4 ポイント·投稿者 decorner·先月·0 コメント
decorner·先月·議論I think it's referring to this article: https://zamin.uz/en/technology/205592-ai-memory-proves-ineff...
decorner·先月·議論Benchmark: https://github.com/tenurehq/precisionmembench - 89 cases measuring retrieval precision indendepently of any LLM. Paper if you want to deep dive on the tech: https://arxiv.org/abs/2605.11325