| title | Using Grounded Theory for Agent Behavior Analysis at Scale |
|---|---|
| authors | Zhuoran Lu, Yangyang Yu, Zhuoyan Li, Yibo Meng, Nan Jiang, Chengxi Zang, Jie Gao, Ziang Xiao |
| date | 2026-08-31 |
| venue | arXiv (preprint, not peer reviewed) |
| id | arXiv:2608.30391 |
| affiliation | Purdue University |
| upvotes | 7 (Hugging Face Daily Papers, 2026-11-01) |
| links | abs pdf hf code |
| abstract | 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. |
cite
@misc{lu2026using,
title={Using Grounded Theory for Agent Behavior Analysis at Scale},
author={Zhuoran Lu and Yangyang Yu and Zhuoyan Li and Yibo Meng and Nan Jiang and Chengxi Zang and Jie Gao and Ziang Xiao},
year={2026},
eprint={2608.30391},
archivePrefix={arXiv},
url={https://arxiv.org/abs/2608.30391}
}