Metraipolis › archive › dc_5003f80aaeb5

Don't Lose the Thread: Empowering Long-Horizon LLM Agents with Cognitive Resource Self-Allocation — arXiv:2610.40350

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.40350
License (as archived)
CC0-1.0
Archived
2026-10-31T06:17:07.683Z
SHA-256 of the archived text
941d1018d1b0e9482f4999476676b5230b931f1d3b30bdd98f7ee9864e0929af
Stable URL of this page
https://metraipolis.com/documento/dc_5003f80aaeb5

How to cite

Don't Lose the Thread: Empowering Long-Horizon LLM Agents with Cognitive Resource Self-Allocation — arXiv:2610.40350 — arXiv (en), https://arxiv.org/abs/2610.40350 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_5003f80aaeb5 (Metraipolis archive dc_5003f80aaeb5, sha256 941d1018d1b0e9482f4999476676b5230b931f1d3b30bdd98f7ee9864e0929af)
BibTeX
@misc{dc5003f80aaeb5,
  title = {Don't Lose the Thread: Empowering Long-Horizon LLM Agents with Cognitive Resource Self-Allocation — arXiv:2610.40350},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.40350},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_5003f80aaeb5; sha256 941d1018d1b0e9482f4999476676b5230b931f1d3b30bdd98f7ee9864e0929af},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.40350. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Agents powered by large language models (LLMs) have demonstrated remarkable progress in solving complex reasoning tasks. However, LLM agents often falter on long-horizon tasks due to cognitive overload, as their working memory becomes cluttered with expanding and irrelevant information, which dilutes their attention and hinders effective planning and reasoning. To mitigate this challenge, we introduce Cognitive Resource Self-Allocation (CORAL), a novel reasoning paradigm that empowers agents to proactively optimize their context. Implemented as an agent-callable working memory management toolset, CORAL allows an agent to checkpoint its task progress and verified facts, and adaptively initiate a new problem-solving episode by resetting cluttered context while preserving only the checkpointed information, enabling it to resume reasoning from a clean state. We further enhance the agent's checkpointing capabilities using a Multi-episode Agentic Reinforced Policy Optimization algorithm. On the GAIA benchmark, CORAL with an 8B model achieves 42.8% average accuracy, with particularly strong gains on the most challenging long-horizon tasks (Level 2 and Level 3). Crucially, these gains come at a substantially lower inference cost: CORAL consumes 6-7$\times$ fewer input tokens than a ReAct agent. It also generalizes to multi-hop Knowledge QA benchmarks and to the considerably harder BrowseComp, where it improves a strong proprietary agent by 8.0 points.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.CL, nuovi lavori del giorno).
Categorie arXiv: Computation and Language (cs.CL).
Pagina del lavoro: https://arxiv.org/abs/2610.40350 — 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 2026-10-31T06:07:23.373Z dalla pagina https://arxiv.org/list/cs.CL/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).