发现大模型推理失败的两种可识别模式。
How Language Models Fail: Token-Level Signatures of Committed and Persistent Reasoning Failures
- 通过令牌级不确定信号，识别出错误锁定和持续不确定两类失败。
- 23组实验中20组验证了预测，证明方法可靠。
- 可用于优化自一致性策略，指导何时该检测或跳过。
语言模型推理中的失败源于可区分的生成过程，并在推理轨迹中留下可识别的标记。我们利用令牌级别的不确定性信号，发现失败主要由两种可实证区分的过程引发：一是‘错误锁定’，模型早期即固定在错误路径上，其关键特征是存在‘承诺点’——此后增加内容反而阻碍失败检测；二是‘持续不确定’，不确定性在整个推理过程中累积，需完整轨迹才能区分成功与失败。这些模式在23个模型-数据集组合中重现，框架的可检验预测在20组中成立，显著高于随机水平。最后，我们证明该失败模式框架对自一致性策略有直接影响，可判断何时使用不确定性信号补充策略，何时可选择性跳过。研究为理解大模型推理失败的可检测性及适配检测策略提供了基础。
原文摘要 · Abstract (English)
Failures in language model reasoning emerge through distinct processes that leave identifiable signatures in the reasoning trace. We characterize these failures using token-level uncertainty signals, finding they arise through two empirically distinguishable processes. The first is committed failure, in which a model locks onto an incorrect reasoning path early in its trace. A central diagnostic signature is the commitment point, beyond which considering additional tokens hurt rather than help failure detection. In the second, persistent uncertainty, uncertainty instead accumulates throughout, and the full trace is needed to best distinguish failing from successful completions. These signatures reproduce across 23 model-dataset configurations, with the framework's falsifiable predictions holding in 20 of 23 cases, well above chance across both failure modes. Finally, we demonstrate our failure mode framework has direct implications for self-consistency, identifying when uncertainty signals complement it and when it can be selectively skipped. These results offer a foundation for understanding when LLM reasoning failures become detectable and for adapting detection strategies accordingly.
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。