{"slug": "compact-memory-llm-agents-via-online-max-member-clustering-and-atom-aware", "title": "Compact-Memory LLM Agents via Online Max-Member Clustering and Atom-Aware Packing", "summary": "Researchers introduced RSM-full, an online clustered-memory pipeline for long-horizon LLM deployments, achieving 83% of Full-Context quality at 32% of the token cost on the AMA-Bench benchmark at a 4k budget. The method outperformed Online K-Means by +3.5–6.0 percentage points across the ~2.6k–5k token regime and beat Budget-RAG by +0.69 pp on the RealMem benchmark, with gains driven by its cosine-gated max-member merge rule and atom-aware grouped context packer.", "body_md": "arXiv:2609.04915v1 Announce Type: new \nAbstract: Many long-horizon LLM deployments face tight prompt budgets: latency, cost, and context limits make full-context prompting impractical as interaction length grows. The key question is then not raw recall alone, but which memory design gives the best quality--token trade-off in the compact-memory regime. We present \\textbf{RSM-full}, an online clustered-memory pipeline designed for a strong quality--token Pareto point.\n  RSM-full combines two design choices: a cosine-gated \\emph{max-member merge} write rule and an atom-aware grouped context packer. On AMA-Bench, our primary compact-memory benchmark, it reaches $83%$ of Full-Context quality at $32%$ of the token cost at a $4$k budget; under four-seed averaging it beats the closest streaming-clustered baseline (Online K-Means) by $+3.5$--$6.0$,pp ($p{<}.001$) across the whole ${\\sim}2.6$k--${\\sim}5$k regime. Three-seed ablations show most of this gain comes from the merge rule ($+5.7$,pp over Online K-Means and matched-$\\tau$ DP-means) and the grouped packer ($+5.0$,pp over flat concatenation).\n  The pattern reproduces on RealMem, an independent long-horizon persona-memory benchmark: RSM-full improves on Budget-RAG ($+0.69$,pp, $p{=}.006$), is on par with BM25-RAG (paired $\\Delta{=}{+}0.27$,pp, $p{=}.47$; we do \\emph{not} claim BM25 equivalence in the equivalence-test sense), and significantly outperforms Streaming-Proto ($+2.97$,pp) and the closest reproduced 2025 agentic-memory baseline A-MEM ($+1.65$,pp, $p{<}.001$). Across benchmarks the message is consistent: under tight budgets, compact-memory performance is driven mainly by how streaming memories are merged and how retrieved content is assembled.\n  Overall, RSM-full is most useful when answeroughly $2k$--$5k$ prompt tokens, where itdefines a strong compact-memory Pareto point; higher-token baselines remain stronger outside this regime.", "url": "https://wpnews.pro/news/compact-memory-llm-agents-via-online-max-member-clustering-and-atom-aware", "canonical_source": "https://www.machinebrief.com/news/compact-memory-llm-agents-via-online-max-member-clustering-a-r0n9", "published_at": "2026-09-07 04:00:00+00:00", "updated_at": "2026-09-07 08:56:16.981992+00:00", "lang": "en", "topics": ["large-language-models", "artificial-intelligence", "ai-research"], "entities": ["RSM-full", "AMA-Bench", "RealMem", "Online K-Means", "Budget-RAG", "BM25-RAG", "Streaming-Proto", "A-MEM"], "alternates": {"html": "https://wpnews.pro/news/compact-memory-llm-agents-via-online-max-member-clustering-and-atom-aware", "markdown": "https://wpnews.pro/news/compact-memory-llm-agents-via-online-max-member-clustering-and-atom-aware.md", "text": "https://wpnews.pro/news/compact-memory-llm-agents-via-online-max-member-clustering-and-atom-aware.txt", "jsonld": "https://wpnews.pro/news/compact-memory-llm-agents-via-online-max-member-clustering-and-atom-aware.jsonld"}}