PulseAugur
中
实时 05:05:36
English(EN) Knowing When to Quit: Diagnosing and Training LLMs to Abort Futile Reasoning

新方法训练大型语言模型拒绝无效推理

研究人员在大型语言模型中发现了一种称为“无效推理”的现象,即模型在超出其能力的任务上生成冗长、复杂但最终不正确的推导。这通常会导致听起来合理但错误的输出,从而误导用户。为了解决这个问题,开发了一种名为 CaRL(能力对齐强化学习)的新方法。CaRL 使用奖励塑造来鼓励模型拒绝不可能的任务,并使用事后拒绝增强来训练模型识别和拒绝无效推理,从而在不牺牲效用的情况下使模型行为与其实际能力保持一致。 AI

影响 这项研究可能带来更可靠的大型语言模型,避免在复杂任务上生成误导性信息。

排序理由 详细介绍大型语言模型训练新方法的学术论文。[lever_c_demoted from research: ic=1 ai=1.0]

在 arXiv cs.CL 阅读 →

AI 生成摘要 · Google Gemini · 来自 1 个来源。 我们如何撰写摘要 →

新方法训练大型语言模型拒绝无效推理

本文如何被排名

Signal score
0 / 100
Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket
Tool
详细介绍大型语言模型训练新方法的学术论文。[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, safety
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
58 days old
Aged out of breaking-news scoring windows; ranking reflects the durable signal from the full source set.

完整方法见我们的编辑标准。

报道来源 [1]

  1. arXiv cs.CL TIER_1 English(EN) · Xinyan Guan, Jiali Zeng, Chunlei Xin, Yaojie Lu, Hongyu Lin, Xianpei Han, Le Sun, Fandong Meng ·

    何时罢手:诊断和训练大型语言模型中止无效推理

    arXiv:2607.29211v1 Announce Type: new Abstract: Large language models generate computationally expensive yet semantically void reasoning on beyond-capability tasks, creating risks where plausible-sounding but incorrect derivations mislead users. We characterize this \textit{futil…