近期研究论文探讨了叙事结构和经验抽象如何影响大型语言模型（LLM）的行为。一项研究表明，LLM会从训练数据中继承叙事模式，这可能带来潜在的对齐风险和意外行为。另一篇论文证明，LLM可以从其解决问题的过程中提取并利用自然语言抽象，从而提高推理任务的性能。第三篇论文强调，任务的叙事框架，而不仅仅是指定的角色，会显著影响LLM的行为，有时会对任务成功产生负面影响。最后，研究表明LLM可能在完全推理之前就承诺答案，即使答案与任务前提相矛盾，初步证据表明这种预先承诺可以在激活层面检测到。 AI
影响 这些研究表明，理解和控制LLM的行为需要解决叙事影响和潜在的预先承诺偏见，这会影响AI系统的开发和对齐方式。
排序理由 该集群包含多篇在arXiv上发表的关于LLM行为和能力的学术论文。
AI 生成摘要 · Google Gemini · 来自 4 个来源。 我们如何撰写摘要 →
arXiv:2607.20449v1 Announce Type: cross Abstract: LLMs are trained predominantly on human-authored text, yet the structural and narrative conventions embedded in that text are rarely examined as a source of systematic behavioral influence, or as a governance risk in deployed syst…
arXiv:2607.20372v1 Announce Type: new Abstract: Humans distill experience into reusable abstractions, e.g., strategies and cautionary reminders, and apply them to gradually solve problems more effectively. We study whether Large Language Models (LLMs) can similarly benefit from s…
arXiv:2607.18566v1 Announce Type: cross Abstract: Persona prompting is widely used to steer LLM agent behavior, yet the narrative framing of a task can matter more than the assigned persona. We isolate this effect through structural isomorphism, constructing three text-based inve…
arXiv:2607.16451v1 Announce Type: cross Abstract: Chat models sometimes commit to an answer and then produce reasoning that justifies it rather than deriving it -- even when the answer contradicts a task premise. We study a minimal probe: "I want to wash my car. The car wash is 1…