Explainable Statute Prediction via Attention-based Model and LLM Prompting
Sachin Pawar, Girish Keshav Palshikar, Anindita Sinha Banerjee, Nitin Ramrakhiyani, Basit Ali · 2025
Read Explainable Statute Prediction via Attention-based Model and LLM Prompting by Sachin Pawar, Girish Keshav Palshikar, Anindita Sinha Banerjee, Nitin Ramrakhiyani, Basit Ali (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2512-21902.txt.
Introduction
Sachin Pawar, Girish Keshav Palshikar, and their co-authors build systems that read a legal case description and predict which statutes apply, with explanations a lawyer could trust. Statute prediction supports assistants and question-answering tools, but predictions are hard to accept in a profession built on reasons.
They propose two techniques. The first, AoS, trains a model that attends over the sentences of the case description; the sentences it highlights double as the explanation. The second, LLMPrompt, asks a large language model in zero-shot fashion, with chain-of-thought prompting, to predict a statute and explain its relevance. The authors compare both against strong baselines on two datasets. Explanation quality is judged two ways: automatically, through counterfactual tests, and by evaluators. The comparison maps a trade-off between a compact, trainable model with built-in explanations and a prompted model that needs no training data.
Catalog record
| CREATOR | Sachin Pawar, Girish Keshav Palshikar, Anindita Sinha Banerjee, Nitin Ramrakhiyani, Basit Ali |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 85 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2512-21902
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Explainable Statute Prediction via Attention-based Model and LLM Prompting Sachin Pawar, Girish Keshav Palshikar, Anindita Sinha Banerjee, Nitin Ramrakhiyani, Basit Ali Abstract In this paper, we explore the problem of automatic statute prediction where for a given case description, a subset of relevant statutes are to be predicted. Here, the term statute refers to a section, a sub-section, or an article of any specific Act. Addressing this problem would be useful in several applications such as AI-assistant for lawyers and legal question answering system. For better user acceptance of such Legal AI systems, we believe the predictions should also be accompanied by human understandable explanations . We propose two techniques for addressing this problem of statute prediction with explanations – (i) AoS ( A ttention- o ver- S entences) which uses attention over sentences in a case description to predict statutes relevant for it and (ii) LLMPrompt which prompts an LLM to predict as well as explain relevance of a certain statute. AoS uses smaller language models, specifically sentence transformers and is trained in a supervised manner whereas LLMPrompt uses larger language models in a zero-shot manner and explores both standard as well as Chain-of-Thought (CoT) prompting techniques. Both these models produce explanations for their predictions in human understandable forms. We compare statute prediction performance of both the proposed techniques with each other as well as with a set of competent baselines, across two popular datasets. Also, we evaluate the quality of the generated explanations through an automated counter-factual manner as well as through human evaluation. keywords Statute Prediction, Explainable AI, Large Language Models † † equal-contributors: Work done while working at TCS Research. † † equal-contributors: Work done while working at TCS Research. 1 Introduction The task of legal statute prediction ( 15 ; 7 ; 20 ) has been an active research area in recent times. In this task, a case description (detailing the facts about the case) is given as input and a set of relevant statutes to the case is expected as an output. In this paper, we use the term statute to refer to a section, a sub-section, or an article of any specific Act (e.g., Indian Penal Code, 1860 ) or convention (e.g., European Convention on Human Rights ). Table 1 shows some examples of statutes along with their textual contents . Here, we use the term statute content for the actual rule or regulation (textual content) corresponding to a statute. Table 2 shows an example of a case description along with a set of statutes relevant to that case. Table 1: Examples of statutes and their textual contents From Indian Penal Code, 1860 (IPC): Section 148 : Whoever is guilty of rioting, being armed with a deadly weapon or with anything which, used as a weapon of offence, is likely to cause death, shall be punished with imprisonment of either description for a term which may extend to three years, or with fine, or with both. Section 465 : Whoever commits forgery shall be punished with imprisonment of either description for a term which may extend to two years, or with fine, or with both. Section 467 : Whoever forges a document which purports to be a valuable security or a will, or an authority to adopt a son, or which purports to give authority to any person to make or transfer any valuable security, or to receive the principal, interest or dividends thereon, or to receive or deliver any money, movable property, or valuable security, or any document purporting to be an acquittance or receipt acknowledging the payment of money, or an acquittance or receipt for the delivery of any movable property or valuable security, shall be punished with imprisonment for life, or with imprisonment of either description for a term which may extend to ten years, and shall also be liable to fine. Section 471 : Whoever fraudulently or dishonestly uses as genuine any document or electronic record which he knows or has reason to believe to be a forged document or electronic record, shall be punished in the same manner as if he had forged such document or electronic record. From European Convention on Human Rights (ECHR): Article 2 : Right to life. 1. Everyone’s right to life shall be protected by law. No one shall be deprived of his life intentionally save in the execution of a sentence of a court following his conviction of a crime for which this penalty is provided by law. 2. Deprivation of life shall not be regarded as inflicted in contravention of this Article when it results from the use of force which is no more than absolutely necessary: (a) in defence of any person from unlawful violence; (b) in order to effect a lawful arrest or to prevent the escape of a person lawfully detained Article 3 : Prohibition of torture. No one shall be subjected to torture or to inhuman or degrading treatment or punishment. Article 8 : Right to respect for private and family life. 1. Everyone has the right to respect for his private and family life, his home and his correspondence. 2. There shall be no interference by a public authority with the exercise of this right except such as is in accordance with the law and is necessary in a democratic society in the interests of national security, public safety or the economic well-being of the country, for the prevention of disorder or crime, for the protection of health or morals, or for the protection of the rights and freedoms of others. Article 14 : Prohibition of discrimination. The enjoyment of the rights and freedoms set forth in this Convention shall be secured without discrimination on any ground such as sex, race, colour, language, religion, political or other opinion, national or social origin, association with a national minority, property, birth or other status. 1.1 Motivation The task of statute prediction can be very useful in practice in multiple applications. For example, consider a lawyer assistant system where a set of relevant statutes are predicted for a case along with explanations. This set of statutes may define the legal scope of the case and any lawyer would find it very useful in deciding future course of actions and/or arguments based on it. For any example case like the one in Table 2 , a lawyer may ask questions to this system such as “ Which part of the crime scenario in the case is most closely related to Section 471? ” and “ Why is Section 467 NOT relevant for the case? ”. Similarly, even for a common person, statute prediction would be useful as part of a legal question answering (QA) system where laypersons can ask the QA system about which different statutes can be used for or against them in their specific case by simply describing the case in natural language. Example of a question can be “ I am a married woman and I was regularly beaten by my husband and in-laws. For filing a case, which sections of IPC are most closely relevant here? ”. Moreover, this task of statute prediction is quite general in nature in the sense that the statutes can be