arXiv cs.CL· Li Zhang, Kevin Ashley·· 13 小时前AI 评分29
法律因素存在于 LLM 的哪些位置？研究揭示 CATO 因子在模型内部的线性可解码表征
Where Do Legal Factors Live in LLMs?
AI 导读
研究检验了 LLM 是否形成对应全部 26 个 CATO 法律因子的内部表征，在 Qwen3-1.7B、Qwen3-8B 和 Gemma-2 2B 三个开源模型中，26 个因子均可线性解码，最佳层测试 AUC 为 0.86 至 1.00。单特征稀疏读出覆盖完整因子目录，5 个选定特征接近全向量读出性能，研究共报告 78 个模型特定的层-特征坐标。
正文 · 原文
Abstract:In case-based legal reasoning, legal professionals identify legally significant fact patterns in a new dispute and compare them with those of decided cases. In AI & Law, HYPO dimensions and CATO factors represent these patterns symbolically, organizing expert knowledge distilled from case narratives as explicit, structured, interpretable, and auditable legal concepts and relations that support case comparison, legal argument, and precedent-based reasoning. Large language models (LLMs) are increasingly used for legal tasks, yet their internal organization of legal knowledge remains difficult to trace. We investigate whether LLMs form internal representations corresponding to all 26 CATO factors. Two research questions examine linear decodability and sparse localization with sparse autoencoders. Experiments cover three open models, Qwen3-1.7B, Qwen3-8B, and Gemma-2 2B, supported by cross-model comparison and human audit of controlled natural-language passages. Results show that all 26 factors are linearly decodable in all three models. Single-feature sparse readouts cover the complete factor catalog, with best-layer test AUCs ranging from 0.86 to 1.00. Five selected features approach full-vector readout performance. We report 78 model-specific layer-feature coordinates, one for each factor in each model. Together, these results show that information corresponding to expert-defined legal factors can be read from the internal activations of LLMs processing legal narratives. By locating these readouts, our study connects legal language and the symbolic abstractions of HYPO and CATO to learned internal representations.
| Comments: | 12 pages, 3 figures, 2 tables. Accepted at JURIX 2026, the 39th International Conference on Legal Knowledge and Information Systems |
| Subjects: | Computation and Language (cs.CL) |
| MSC classes: | 68T50 |
| ACM classes: | I.2.7; I.2.4 |
| Cite as: | arXiv:2610.28799 [cs.CL] |
| (or arXiv:2610.28799v1 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2610.28799 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Li Zhang [view email]
[v1]
Tue, 20 Oct 2026 18:24:37 UTC (585 KB)
来源：arXiv cs.CL · arxiv.org