aditisinghh17 / LegalRAG-Evaluation Public

Notifications You must be signed in to change notification settings
Fork 1
Star 1

This repository contains evaluations of LLM-based Retrieval-Augmented Generation (RAG) pipelines tailored for the legal domain. It currently features RAG with three different LLMs: Mistral, OpenAI, and Phi3:Medium.

1 star 1 fork Branches Tags Activity

Notifications

Name		Name	Last commit message	Last commit date
Latest commit History 54 Commits
Output-LLMs		Output-LLMs
datasets_used		datasets_used
.gitignore		.gitignore
LICENSE		LICENSE
RAG_OPENAI (2).ipynb		RAG_OPENAI (2).ipynb
RAG_Phi3_Medium.ipynb		RAG_Phi3_Medium.ipynb
RAG_mistral.ipynb		RAG_mistral.ipynb
README.md		README.md

Repository files navigation

Assessment of Legal Domain-Specific RAG Pipelines Using Diverse LLMs

COMPONENTS OF RAG SYSTEM

LLMs Used : ChatGPT Turbo-3.5, Phi3:Medium, Mistral
Embeddings : OpenAI
Vector Store : Pinecone
LangChain
Chunk size = 1000, chunk overlap = 200

DATASET

Knowledge Base : https://huggingface.co/datasets/ninadn/indian-legal
Q/A Set (generated synthetically with GPT-3.5 Turbo) : Q/A Dataset

About

This repository contains evaluations of LLM-based Retrieval-Augmented Generation (RAG) pipelines tailored for the legal domain. It currently features RAG with three different LLMs: Mistral, OpenAI, and Phi3:Medium.

Report repository

Releases

No releases published

Packages

No packages published

Contributors 2

Languages

Jupyter Notebook 100.0%