按年被提到
20252025 · 1 次2026
AI 安全、伦理与社会影响›大模型与人类行为·arXiv:2502.06258 · cs.CL

Emergent Response Planning in LLMs

Zhichen Dong, Zhanhui Zhou, Zhixuan Liu, Chao Yang, Chaochao Lu · 2025
开始阅读 ↗PDFalphaXiv 讨论
被提及 · 2 人首次 2025-02-12 · 最近 2026-02-06
2不同的人
2活跃月份
2讨论串
4被引(OpenAlex)
arXiv 摘要(候选层不写导读)

In this work, we argue that large language models (LLMs), though trained to predict only the next token, exhibit emergent planning behaviors: $\textbf{their hidden representations encode future outputs beyond the next token}$. Through simple probing, we demonstrate that LLM prompt representations encode global attributes of their entire responses, including $\textit{structure attributes}$ (e.g., response length, reasoning steps), $\textit{content attributes}$ (e.g., character choices in storywriting, multiple-choice answers at the end of response), and $\textit{behavior attributes}$ (e.g., answer confidence, factual consistency). In addition to identifying response planning, we explore how it scales with model size across tasks and how it evolves during generation. The findings that LLMs plan ahead for the future in their hidden representations suggest potential applications for improving transparency and generation control.

候选层:至少 2 个不同的人、在 2 个不同月份提到过,还没到主书架的 3 人 3 个月。没有做立场标注,所以只写“被提及”。

HN 书评

没有评论直接链到它;下面是它被提交的讨论串里回复最多的评论(不计入统计)

讨论串里没有 ≥100 字的评论,只有提交记录。

被提交为 story · 2 次