English(EN) When Answer Decodability Is Not a Decision: Separating Final-Answer and Imminent-Emission Signals in Chain-of-Thought

新的探测方法区分大型语言模型中迫在眉睫的输出和最终答案

作者 PulseAugur 编辑部 · [1 个来源] · 2026-11-09 05:14

研究人员开发了一种名为“发射控制双目标探测”（ECDP）的新方法，以更好地理解语言模型内的决策过程。该技术区分了模型即将输出的信号（它即将说什么）与其最终答案信号。使用Qwen3.5-4B和Llama-3.1-8B等模型在特定任务上的实验表明，即将输出的信号可能比最终答案信号更强，这挑战了强答案读出总是表明已做出承诺决策的假设。ECDP的有效性取决于所使用的模型、任务和探测协议。 AI

影响提供了对大型语言模型如何生成响应的更细致的理解，有可能提高可解释性和控制性。

排序理由详细介绍分析大型语言模型内部状态新方法的论文。[lever_c_demoted from research: ic=1 ai=1.0]

在 Hugging Face Daily Papers 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: 1 / 100; Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket: Tool; 详细介绍分析大型语言模型内部状态新方法的论文。[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: 1 days old; Coverage has settled into its steady-state source set.

完整方法见我们的编辑标准。

报道来源 [1]

Hugging Face Daily Papers TIER_1 English(EN) · 2026-11-09 05:14

当答案可解码性并非决定因素时：在思维链中分离最终答案和迫在眉睫的输出信号

Answer probes applied to chain-of-thought hidden states are often read as evidence that a language model has already decided its final response. This reading overlooks a competing signal, the candidate the model is about to mention. We introduce emission-controlled dual-target pr…

本文如何被排名

报道来源 [1]

当答案可解码性并非决定因素时：在思维链中分离最终答案和迫在眉睫的输出信号

相关实体

相关话题