Metraipolis › archive › dc_93af9cd49c04
Persistent Memory in Multi-Agent LLM Inference: What It Costs, What It Buys, and When You Can Tell — arXiv:2610.07782
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.07782
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-07T11:19:06.569Z
- SHA-256 of the archived text
fc7cc2246f9c67d909af66b1a15d03a19c693d976bce4b6749380134903026d2- Stable URL of this page
https://metraipolis.com/documento/dc_93af9cd49c04
How to cite
Persistent Memory in Multi-Agent LLM Inference: What It Costs, What It Buys, and When You Can Tell — arXiv:2610.07782 — arXiv (en), https://arxiv.org/abs/2610.07782 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_93af9cd49c04 (Metraipolis archive dc_93af9cd49c04, sha256 fc7cc2246f9c67d909af66b1a15d03a19c693d976bce4b6749380134903026d2)
BibTeX
@misc{dc93af9cd49c04,
title = {Persistent Memory in Multi-Agent LLM Inference: What It Costs, What It Buys, and When You Can Tell — arXiv:2610.07782},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.07782},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_93af9cd49c04; sha256 fc7cc2246f9c67d909af66b1a15d03a19c693d976bce4b6749380134903026d2},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.07782. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Decomposing long-context inference across cooperating agents bounds the active KV cache per call rather than total evidence, which matters when KV-cache memory binds. Many such systems add a persistent tier storing and recalling reasoning traces, usually validated by an ablation reporting an accuracy gain. We measure both on one three-tier agent architecture. Decomposition delivers: peak KV working set of 14.3 MiB per query against 35.5 and 35.3 MiB for single-pass and retrieval-augmented baselines. The persistent tier does not: across eight controlled dataset pairs at n=100 per arm it costs +0.368 MiB [+0.167, +0.590] of peak cache and produces no detectable accuracy change (+0.015, 95% CI [-0.011, +0.046]). We argue the null is structural: single-question benchmarks supply each item with its own evidence and score it independently, and correctness requires resetting stored traces between conditions, so recall has nothing informative to retrieve. Reaching it took four measurement corrections -- three inflating the apparent benefit, the fourth making an effect that size look resolvable -- none visible in the results table. We give the conditions an agent-memory ablation must satisfy and detection procedures that need no knowledge of the specific defect. Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno). Categorie arXiv: Artificial Intelligence (cs.AI) ; Computation and Language (cs.CL); Distributed, Parallel, and Cluster Computing (cs.DC); Machine Learning (cs.LG). Pagina del lavoro: https://arxiv.org/abs/2610.07782 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-10-07T11:15:24.552Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).