---
id: 20260611-T0-12
title: "研究发现：多智能体辩论中谎报自信度的识别方法"
title_en: "Detecting Confident Liars in Multi-Agent Debates"
url: https://ai.daily.yangsir.net/daily/20260611-T0-12
issue_date: 2026-06-11
publish_date: 2026-06-10T04:00:00.000Z
category: research
source_name: "arXiv cs.CL (NLP)"
source_url: https://arxiv.org/abs/2606.10296
---

# 研究发现：多智能体辩论中谎报自信度的识别方法

论文提出诊断多智能体辩论中虚假自信的方法。传统评估只关注最终答案正确性，忽略推理质量。新方法通过分析对数概率和LLM作为评判员，识别自信但错误的论证。实验证明该方法能有效识别'自信说谎者'，提升辩论系统可靠性。代码已开源。

## English Version

**Detecting Confident Liars in Multi-Agent Debates**

New research introduces a method to detect 'confident liars' in multi-agent debates. Traditional evaluations only check final answers, ignoring reasoning quality. The new approach uses log-probabilities and LLM-as-judge to identify confident but incorrect arguments. Experiments show it effectively detects false confidence, improving debate system reliability.

---

**来源**：[arXiv cs.CL (NLP)](https://arxiv.org/abs/2606.10296)

**详情页**：https://ai.daily.yangsir.net/daily/20260611-T0-12

---

*智语观潮 · Daily — https://ai.daily.yangsir.net/llms.txt*