Metraipolis › archive › dc_38ca18370c0d
HasMem: Hard-Origin Adaptively Softened Memory for Long-Term LLM Agents — arXiv:2609.30797
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.30797
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-08T01:38:20.906Z
- SHA-256 of the archived text
78be15516c8a2a71ed2e24d66a06ce5f742f80f0c13a3e12042b68d0137b8fb2- Stable URL of this page
https://metraipolis.com/documento/dc_38ca18370c0d
How to cite
HasMem: Hard-Origin Adaptively Softened Memory for Long-Term LLM Agents — arXiv:2609.30797 — arXiv (en), https://arxiv.org/abs/2609.30797 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_38ca18370c0d (Metraipolis archive dc_38ca18370c0d, sha256 78be15516c8a2a71ed2e24d66a06ce5f742f80f0c13a3e12042b68d0137b8fb2)
BibTeX
@misc{dc38ca18370c0d,
title = {HasMem: Hard-Origin Adaptively Softened Memory for Long-Term LLM Agents — arXiv:2609.30797},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.30797},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_38ca18370c0d; sha256 78be15516c8a2a71ed2e24d66a06ce5f742f80f0c13a3e12042b68d0137b8fb2},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.30797. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Text-based memory and context compression support reuse of past interactions. Resizing continuous memory changes the input to a frozen LLM, coupling capacity allocation with readout. We propose Hard-Origin Adaptively Softened Memory (HasMem). Frozen hard-prompt embeddings provide a verifiable initial state. A controller adjusts memory widths, a Writer re-encodes resized entries, and Reader and Global provide readout adaptation and cross-turn state. On all $535$ questions in a reconstruction probe derived from the Multi-Session Chat (MSC) development split, the main configuration achieves lexical F1 of $95.3$ ($+4.4$ percentage points) at $93.6\%$ of the hard reference's framed memory positions. With approximately matched per-question target body budgets, six configurations at mean per-entry retention around $0.83$--$0.91$ exceed rule-based re-encoding by $8.0$--$23.6$ exact-match (EM) percentage points. With fixed model parameters and rule target width ratio $0.75$, Global's EM gain passes a user-level exact paired test with Bonferroni correction over eight comparisons. On all $500$ LongMemEval-S questions, local lexical F1 rises from the hard reference's $3.4$ to $8.9$, and answer negative log-likelihood (NLL) falls from $12.257$ to $5.274$. F1 gains accompany lower EM on both evaluations. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-28 in cs.AI). Categorie arXiv: Artificial Intelligence (cs.AI). Pagina del lavoro: https://arxiv.org/abs/2609.30797 — 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 2026-10-08T01:12:14.288Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-28?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).