English(EN) ReFold: Training-Free Reversible Inter-Turn Context Folding for Long-Horizon Agents

ReFold方法在不损失性能的情况下将AI智能体上下文成本降低3.4倍

作者 PulseAugur 编辑部 · [1 个来源] · 2026-10-07 04:00

研究人员开发了ReFold，一种新颖的、无需训练即可管理长时程AI智能体上下文的方法。该方法通过移除冗余信息和总结已完成的回合来压缩渲染的上下文，而不会永久丢弃内容。ReFold作为ReAct风格智能体的即插即用层运行，显著降低了代币消耗、KV缓存内存和推理成本。评估表明，它可以在不影响任务成功率的情况下将代币使用量最多降低2.5倍，并将内存使用量减半，同时还能加速推理并减少请求排队延迟。 AI

影响降低了长时程AI智能体的计算成本并提高了效率，可能支持更复杂和更长的任务。

排序理由该集群包含一篇详细介绍AI智能体新方法的论文。[lever_c_demoted from research: ic=1 ai=1.0]

在 arXiv cs.AI 阅读 →

ReAct

AI 生成摘要 · Google Gemini · 来自 1 个来源。我们如何撰写摘要 →

本文如何被排名

Signal score: 15 / 100; Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket: Tool; 该集群包含一篇详细介绍AI智能体新方法的论文。[lever_c_demoted from research: ic=1 ai=1.0]
Source corroboration: Single-source cluster; Only one publisher covered this so far. Single-source stories can still rank when the publisher is high-authority, but they lack cross-source corroboration.
Topics: paper, infra; Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance: High; Clearly on-topic for AI-industry coverage.
Story freshness: Breaking (< 6h); Fresh story with cross-source coverage still developing. Ranking may shift as more sources report.

完整方法见我们的编辑标准。

报道来源 [1]

arXiv cs.AI TIER_1 English(EN) · Yupeng Su, Jiayi Tian, Zheng Zhang, Souvik Kundu · 2026-10-07 04:00

ReFold：用于长视域智能体的无训练可逆跨匝上下文折叠

arXiv:2610.07863v1 Announce Type: cross Abstract: Long-horizon LLM agents act on an append-only interaction history that is re-sent to the model at every step, so the context and its cost grow with steps until the sessions exceed the context window. Existing methods manage the co…

本文如何被排名

报道来源 [1]

ReFold：用于长视域智能体的无训练可逆跨匝上下文折叠

相关实体

相关话题