PulseAugur
实时 09:50:46

新的AI挑战测试LLM的心理理论能力,Gemini3-Pro和GPT-5.4表现不佳

一篇新的研究论文介绍了一个名为“用于引导信念的心理理论”(ToM-SB)的挑战,旨在测试大型语言模型理解和操纵他人信念的能力,类似于心理理论。研究发现,像Gemini3-Pro和GPT-5.4这样的先进模型在此任务上表现不佳,尤其是在攻击者拥有部分先验知识的情况下。为了解决这个问题,研究人员使用强化学习训练了AI双重代理,证明奖励信念操纵和对攻击者心理状态的理解可以显著提高性能。 AI

影响 强调了在LLM中改进心理理论能力以实现更安全、更复杂的交互的必要性。

排序理由 详细介绍新基准和LLM发现的研究论文。[lever_c_demoted from research: ic=1 ai=1.0]

在 arXiv cs.AI 阅读 →

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

新的AI挑战测试LLM的心理理论能力,Gemini3-Pro和GPT-5.4表现不佳

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
详细介绍新基准和LLM发现的研究论文。[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
53 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.AI TIER_1 English(EN) · Hanqi Xiao, Vaidehi Patil, Zaid Khan, Hyunji Lee, Elias Stengel-Eskin, Mohit Bansal ·

    配合演出:通过心智理论学习双重代理防御者以进行信念引导

    arXiv:2604.11666v2 Announce Type: replace-cross Abstract: As large language models (LLMs) become the engine behind conversational systems, their ability to reason about the intentions and states of their dialogue partners (i.e., form and use a theory-of-mind, or ToM) becomes incr…