ragR
Provides tools for document ingestion, embedding storage, retrieval-augmented generation (RAG), and evaluation of question-answering systems. The package includes an R-native vector store, wrappers for OpenAI embedding and chat-completion application programming interfaces (APIs), question-answering logging utilities, and large language model (LLM)-based evaluation metrics for context precision, context recall, answer relevance, and faithfulness. These metrics are based on the Retrieval-Augmented Generation Assessment (RAGAS) framework. The retrieval-augmented generation methodology is described by Lewis et al. (2020) "Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks" <doi:10.48550/arXiv.2005.11401>. The evaluation metrics are based on Es et al. (2024) "RAGAS: Automated Evaluation of Retrieval Augmented Generation" <doi:10.18653/v1/2024.eacl-demo.16>.
This page was built for software: ragR
