PROSLEX: Expert-Annotated Dataset for Legal Statute Prediction in Indian Judiciary
Researchers have introduced PROSLEX, a novel dataset designed to advance Legal Statute Prediction (LSP) by integrating legal reasoning into the process. The dataset comprises 1,623 expert-annotated legal documents from the Indian context, each paired with statute predictions and detailed explanations, totaling 7,450 explanations. This initiative addresses a critical gap in current LSP research, which primarily focuses on accuracy metrics without considering the explainability and justifiability required in judicial contexts. By capturing underlying legal reasoning, PROSLEX aims to enhance the use of Large Language Models (LLMs) in legal applications, making them more aligned with judicial needs. The dataset is expected to facilitate research in natural language processing and information retrieval, particularly in multi-label classification tasks. The announcement was made on arXiv under the identifier 2608.08830, with the paper detailing the dataset's construction and potential applications.
Key facts
- PROSLEX is a new dataset for Legal Statute Prediction (LSP).
- It includes 1,623 expert-annotated legal documents from India.
- The dataset contains 7,450 explanations capturing legal reasoning.
- It addresses the lack of explainability in current LSP methods.
- The work is published on arXiv with ID 2608.08830.
- The dataset is designed for multi-label classification tasks.
- It aims to improve LLM-based statute prediction with legal reasoning.
- The research focuses on the Indian judicial context.
Entities
Institutions
- arXiv
Locations
- India