给AI行为做田野调查：六十年老方法管用
研究AI智能体行为，过去靠人眼看轨迹或预设分类器，但任务一长、一陌生，预设就失灵。这篇把社会学用了六十年的「扎根理论」搬来：不预设框架，让AI自己从海量轨迹里一层层归纳出行为类别，直到新信息不再出现为止。在六个轨迹数据集上，它自动生成的分类能覆盖人类标注的73%到91%的失败模式，还额外发现了人类漏掉的规律。它不是你明天能直接用的工具，但提示了一个方向：研究AI行为，或许该像研究人类社会一样，先观察、再归纳，而不是先定标签。
📄 原文摘要(英文)
Understanding agent behavior requires methods that scale to thousands of trajectories and surface new patterns in long, often unfamiliar tasks where pre-built classifiers fall short. We propose to bring grounded theory into agent trajectory analysis: a six-decade-old qualitative method from the social sciences, with a principled saturation criterion and an auditable trail from data to theory. We propose AutoTraceGT (Automated Trace analysis through Grounded Theory), the first multi-agent pipeline that automates grounded theory on agent trajectories. It iteratively performs open, axial, and theoretical coding until saturation, producing a behavioral taxonomy tailored to each task. Across six trajectory corpora, AutoTraceGT produces codebooks that recover 73-91 percent of the failure modes in human-annotated taxonomies and surface additional patterns that those taxonomies miss. The emergent theoretical narrative aligns with prior expert accounts. Used as a deductive feature space, the codebook outperforms zero-shot and few-shot LLM baselines on downstream failure prediction. These results suggest Grounded Theory offers a scalable analytic tool for ML researchers and agent developers studying what agents actually do.