English(EN) What if LLMs Ate Their Words: Causal History Effects in Multi-Turn Interaction

大型语言模型过去的响应会降低多轮交互性能

作者 PulseAugur 编辑部 · [1 个来源] · 2026-09-10 04:00

一篇新的研究论文探讨了大型语言模型（LLMs）如何受到其自身在多轮对话中过去响应的影响。研究发现，LLM先前的输出会显著降低性能，并且影响因模型和任务而异。通过干预以中和或编辑特定的过去助手响应，表明历史管理而非简单的上下文缩短是提高交互鲁棒性的关键。该研究还确定了与这些行为变化相关的、依赖于任务的内部特征。 AI

影响强调了改进LLM历史管理的需求，以确保对话式AI的性能一致性。

排序理由学术论文，详细介绍了关于LLM行为的新研究发现。[lever_c_demoted from research: ic=1 ai=1.0]

在 arXiv cs.AI 阅读 →

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: 13 / 100; Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket: Tool; 学术论文，详细介绍了关于LLM行为的新研究发现。[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, model release; 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) · Jinnan Li, Zheren Fu, Yue Wang, Jinzhe Li, Yuan Wu, Yi Chang · 2026-09-10 04:00

如果大型语言模型“吃掉”它们说过的话：多轮交互中的因果历史效应

arXiv:2609.05882v1 Announce Type: cross Abstract: Multi-turn interaction creates a feedback process in which an LLM's previous responses become context for later behavior. Prior work shows substantial multi-turn degradation and that assistant-generated history can affect later be…

本文如何被排名

报道来源 [1]

如果大型语言模型“吃掉”它们说过的话：多轮交互中的因果历史效应

相关实体

相关话题