English(EN) Evaluating Confidence-Gated Retrieval with Matched Trajectory Replay

新协议评估语言模型代理中的置信信号

作者 PulseAugur 编辑部 · [1 个来源] · 2026-08-28 04:00

研究人员开发了一种名为匹配轨迹回放的新协议，用于评估语言模型代理如何使用置信信号来决定在回答前是否检索额外信息。该方法控制了候选答案状态和证据点等变量，以比较不同的置信度到行动的映射。使用 Mistral、GPT 和 Qwen 模型在问答数据集上的实验表明，虽然校准可以提高准确性并使承诺风险可解释，但它不一定能估计进一步检索的好处，有时还会降低覆盖率或增加检索使用。 AI

影响这项研究通过改进 AI 代理评估外部信息需求的方式，有望带来更可靠、更可解释的决策。

排序理由该集群包含一篇详细介绍语言模型代理新评估协议的学术论文。[lever_c_demoted from research: ic=1 ai=1.0]

在 arXiv cs.AI 阅读 →

GPT
Qwen

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: 22 / 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语言模型代理新评估协议的学术论文。[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, 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) · Prateek Chhikara · 2026-08-28 04:00

评估带匹配轨迹重放的置信门控检索

arXiv:2608.26846v1 Announce Type: cross Abstract: Interactive language-model agents use confidence signals to decide whether to answer immediately, retrieve additional evidence (from memory or external knowledge), or defer. Yet confidence is usually evaluated in isolation, withou…

本文如何被排名

报道来源 [1]

评估带匹配轨迹重放的置信门控检索

相关实体

相关话题