A Two-Stage Architecture for NDA Analysis: LLM-based Segmentation and Transformer-based Clause Classification

TL;DR

This paper introduces a two-stage AI architecture using LLMs and transformers to automate NDA document segmentation and clause classification, improving efficiency and accuracy in legal document analysis.

Contribution

It presents a novel two-model pipeline combining LLaMA-3.1 and Legal-Roberta-Large for NDA analysis, achieving high segmentation and classification performance.

Findings

Segmentation ROUGE F1 of 0.95

Clause classification weighted F1 of 0.85

Demonstrates effective automation of NDA analysis

Abstract

In business-to-business relations, it is common to establish NonDisclosure Agreements (NDAs). However, these documents exhibit significant variation in format, structure, and writing style, making manual analysis slow and error-prone. We propose an architecture based on LLMs to automate the segmentation and clauses classification within these contracts. We employed two models: LLaMA-3.1-8B-Instruct for NDA segmentation (clause extraction) and a fine-tuned Legal-Roberta-Large for clause classification. In the segmentation task, we achieved a ROUGE F1 of 0.95 +/- 0.0036; for classification, we obtained a weighted F1 of 0.85, demonstrating the feasibility and precision of the approach.