arXiv cs.CL· Haokai Zhao, Jie Gao, Yunze Xiao, Xintao Wang, Weihao Xuan, Aditya Joshi, Mark Dredze, Jen-tse Huang·· 13 小时前AI 评分38
研究提出 A-B-D 方法，从行为轨迹推断 LLM 智能体性格特质
On the Behavioral Traits of LLM Agents
AI 导读
研究者提出 A-B-D 方法，从 345,667 条真实轨迹中自下而上推断 LLM 智能体的行为特质，覆盖 80 个模型、12 项任务和 50 个 harness。经筛选的 79 个特征经因子分析得出 6 个稳定的模型可归因因子，其中 2 个功能性、4 个语言性。这些因子与模型自报的 Big Five 分数仅弱相关，如外向性与 energetic 的 r = 0.07，揭示出“知行差距”。
正文 · 原文
Abstract:Users increasingly describe different AI agents as distinct colleagues to work with. AI personality research aims to quantify such impressions by attributing human- like "traits" to agents. However, existing measures fall short: models' self-reports (S-data) diverge from their actual behavior, while informant ratings from LLM judges (I-data) are costly to scale and cover few everyday scenarios. In this paper, we propose A-B-D to infer traits bottom-up from behavioral data (B-data), namely how agents act on their environment and communicate with users, as recorded in existing trajectories. From 345,667 real-world trajectories spanning 80 models, 12 tasks, and 50 harnesses, we extract 318 candidate features that capture both the actions an agent takes at each step (functional) and the language accompanying them (linguistic). We retain only features that show instance-level stability, cross-task consistency, and model discriminability. Factor analysis of the remaining 79 features uncovers six stable, model-attributable factors, two functional and four linguistic. For example, Claude-Sonnet-4.6 and Kimi-K3 exhibit the most planfulness, whereas GPT-5.5 and GPT-5.6 are the least energetic. Moreover, we quantify the "knowledge-action gap" in the wild: these factors correlate only weakly with self-reported Big Five scores, even for conceptually matched pairs such as extraversion and energetic (r = 0.07, p = 0.58). Our work offers a new lens for understanding AI personality, with implications for users, developers, and researchers from both computer science and social science.
| Comments: | Preprint; working in progress |
| Subjects: | Computation and Language (cs.CL) |
| Cite as: | arXiv:2609.32776 [cs.CL] |
| (or arXiv:2609.32776v2 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2609.32776 arXiv-issued DOI via DataCite |
Submission history
From: Jen-Tse Huang [view email]
[v1]
Sat, 26 Sep 2026 16:51:52 UTC (3,939 KB)
[v2]
Thu, 22 Oct 2026 14:01:08 UTC (4,173 KB)
来源：arXiv cs.CL · arxiv.org