arXiv cs.CL· Md. Rafiul Biswas, George Mikros, Wajdi Zaghouani·· 12 小时前AI 评分26
从上下文视角看压缩：阿拉伯语法律 LLM 的适配、量化与蒸馏评估
Compression Through a Context Lens: Evaluating Adaptation, Quantization, and Distillation in Arabic Legal LLMs
AI 导读
研究基于沙特阿拉伯语法律基准 ArabLegalEval，在带上下文与无上下文两种评测下考察 FANAR-9B、JAIS-8B、Qwen2.5-7B 的领域适配、量化与知识蒸馏效果。
正文 · 原文
Abstract:Model compression can facilitate the deployment of large language models in resource-constrained legal settings, but its effects on domain-specific performance may depend on the evaluation protocol. We investigate this issue using ArabLegalEval, a Saudi Arabic legal benchmark, under both with-context and no-context evaluation. The former provides supporting legal passages, while the latter assesses performance without external context. Removing the context produces accuracy gaps of up to 27 percentage points for the base model. We examine three stages: domain adaptation, quantization, and knowledge distillation across FANAR-9B, JAIS-8B, and Qwen2.5-7B. Four-bit quantization reduces model memory by 64-71%, with changes in no-context MCQ accuracy ranging from -3.2 to +1.6 percentage points. Distillation produces more variable results, with accuracy generally declining as student size decreases. A Qwen2.5-7B student distilled from FANAR achieves 76.6% no-context accuracy, compared with 77.2% for its teacher. Our findings suggest that no-context evaluation provides a useful complement to context-supported testing when assessing the effects of adaptation and compression on Arabic legal language models.
| Comments: | 2 figures, 4 tables |
| Subjects: | Computation and Language (cs.CL) |
| Cite as: | arXiv:2611.07988 [cs.CL] |
| (or arXiv:2611.07988v1 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2611.07988 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Md.Rafiul Biswas Dr. [view email]
[v1]
Thu, 5 Nov 2026 13:15:55 UTC (1,801 KB)
来源：arXiv cs.CL · arxiv.org