Metraipolis › archive › dc_75dd23a72d41

Role-aware Heuristic Episodic Attention for Conversational LLMs — arXiv:2610.00958

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.00958
License (as archived)
CC0-1.0
Archived
2026-10-06T13:54:38.548Z
SHA-256 of the archived text
d2e0cc031f70d459a5a6f29494f66beed3a76996ed96709e066bbc276de217c2
Stable URL of this page
https://metraipolis.com/documento/dc_75dd23a72d41

How to cite

Role-aware Heuristic Episodic Attention for Conversational LLMs — arXiv:2610.00958 — arXiv (en), https://arxiv.org/abs/2610.00958 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_75dd23a72d41 (Metraipolis archive dc_75dd23a72d41, sha256 d2e0cc031f70d459a5a6f29494f66beed3a76996ed96709e066bbc276de217c2)
BibTeX
@misc{dc75dd23a72d41,
  title = {Role-aware Heuristic Episodic Attention for Conversational LLMs — arXiv:2610.00958},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.00958},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_75dd23a72d41; sha256 d2e0cc031f70d459a5a6f29494f66beed3a76996ed96709e066bbc276de217c2},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.00958. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Large language models often lose track of persistent instructions and relevant information as multi-turn conversations grow. We study this cumulative contextual decay through three related failure modes: attention pollution, dilution, and drift. We propose REA (Role-aware Heuristic Episodic Attention), a context-management framework that assigns different persistence and representation policies to instructions and episodic interactions. Instructional Memory retains identified global constraints in a dedicated prefix. Episodic Memory preserves user inputs and compresses model replies, while heuristic retrieval selects raw text, compressed representations, or omission for each historical turn. On Long-MT-Bench+, REA improves the judge score from 6.32 to 7.36 on a 10-point scale, a 16.5% relative gain over the Vanilla baseline, and reduces average latency by 2.91$\times$. Additional evaluations show aggregate gains on three backbones spanning 1.7B-7B parameters and on Chinese and English role-playing tasks. These results support role-aware context management as a practical approach to maintaining conversational continuity and instruction adherence.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-10-02 in cs.CL).
Categorie arXiv: Computation and Language (cs.CL).
Pagina del lavoro: https://arxiv.org/abs/2610.00958 — 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 2026-10-06T12:15:34.663Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-10-02?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).