提出新方法检测大模型推理中的错误传播，提升结论可靠性。
Probabilistic Soundness Guarantees in LLM Reasoning Chains
- 基于已验证前提逐步评估推理，避免早期错误影响判断
- 在四个基准上达到72.1%宏F1，长链推理错误检测率达90.3%
- 提供统计可信保证，适合需要高可靠推理的场景
大型语言模型（LLM）生成的推理链中，初始错误常会传播并削弱最终结论的可信度。现有基于LLM的错误检测方法往往无法发现这些传播错误，因为早期错误会污染下游推理的判断。为此，我们提出自回归推理蕴含稳定性（ARES），一种基于概率的框架，仅依据先前已验证的前提评估每一步推理。该归纳方法为每一步生成细致评分，提供经认证的统计可靠性保障，而非脆弱的二值标签。ARES在四个基准测试中表现最优（宏F1 72.1%，提升8.2点），在极长合成推理链上展现出更强鲁棒性，能有效检测传播错误（F1达90.3%，提升27.6点）。
原文摘要 · Abstract (English)
In reasoning chains generated by large language models (LLMs), initial errors often propagate and undermine the reliability of the final conclusion. Current LLM-based error detection methods often fail to detect propagated errors because earlier errors can corrupt judgments of downstream reasoning. To better detect such errors, we introduce Autoregressive Reasoning Entailment Stability (ARES), a probabilistic framework that evaluates each reasoning step based solely on previously-verified premises. This inductive method yields a nuanced score for each step and provides certified statistical guarantees of its soundness, rather than a brittle binary label. ARES achieves state-of-the-art performance across four benchmarks (72.1% Macro-F1, +8.2 points) and demonstrates superior robustness on very long synthetic reasoning chains, where it excels at detecting propagated errors (90.3% F1, +27.6 points).
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。