Lightweight Domain Adaptation of a Large Language Model for Legal Assistance in the Indian Context

arXiv cs.CL / 5/4/2026

💬 OpinionIdeas & Deep AnalysisTools & Practical UsageModels & Research

共有:

Key Points

The paper proposes Legal Assist AI, an efficient framework to deliver legal assistance in India by addressing the public’s limited access to accurate legal information.
It shows that an 8B-parameter quantized Llama 3.1 model can outperform a much larger 175B-parameter GPT-3.5 Turbo for the legal domain by combining RAG with targeted prompt engineering.
The approach relies on a continually updated, high-quality corpus of 600+ Indian legal documents, including the Constitution and recently enacted laws such as the Bharatiya Nyaya Sanhita (BNS) and Bharatiya Nagarik Suraksha Sanhita (BNSS).
On the All-India Bar Examination (AIBE) benchmark, the system reaches 60.08%, improving over GPT-3.5 Turbo’s 58.72%, suggesting strong practical effectiveness for legal Q&A.
The framework is reported to mitigate hallucinations and introduces a Parameter Efficiency Index (PEI), finding the 8B model is 22× more parameter-efficient than the 175B baseline, supporting the value of smaller domain-adapted models.

Abstract

In India, access to legal assistance for the general public has been observed to have a critical gap, as many citizens are not able to take full advantage of their legal rights due to limited access and awareness of apposite legal information. This paper thus introduces Legal Assist AI, a highly efficient framework designed to provide legal assistance in the Indian domain. The core contribution is a framework demonstrating how a smaller, 8-billion-parameter quantized model (Llama 3.1) can achieve superior domain-specific performance. This effective performance stems from integrating a Retrieval-Augmented Generation (RAG) system with strategic prompt engineering, supported by a high-quality, up to date corpus of more than 600 legal documents. This corpus includes the Indian Constitution and more importantly, the newly enacted Bharatiya Nyaya Sanhita (BNS) and Bharatiya Nagarik Suraksha Sanhita (BNSS) among others. Further, by achieving a score of 60.08\% in the All-India Bar Examination (AIBE) benchmark, the specialized approach based on RAG was found to be highly efficient and effective, improving on the 58.72\% score of the 175-billion parameter GPT-3.5 Turbo. It was also observed that the framework was able to manage and mitigate instances of hallucinations successfully, which is a critical requirement for practical legal applications. A Parameter Efficiency Index (PEI) is also introduced, with the goal of quantifying the superior efficiency that the framework was able to achieve, demonstrating how the 8B model is 22 times more parameter-efficient than the 175B baseline, and hence corroborating the potential of smaller domain-adapted models.