From the course: LLM Evaluations and Grounding Techniques
Join today to access over 25,300 courses taught by industry experts.
Ragas: Evaluation paper
From the course: LLM Evaluations and Grounding Techniques
Ragas: Evaluation paper
- [Narrator] So far, we've been very hands-on, learning to ground LLMs with a focus on coding and real life examples. In this chapter, we'll be going a little bit more theoretical, learning how to analyze five grounding papers that have been impactful in the space. We're going to be starting off with a RAG paper called RAGAS. RAGAS is a paper from 2023 that focuses on the automatic evaluation of retrieval augmented generation. Now there are a few interesting things here. It introduces the concept of measuring the correctness of RAG responses. These are broken down into a few different metrics. If we scroll down to page three, we get a definition of these three papers. Faithfulness, answer relevance, and context relevance. We already talked about faithfulness and how well an LLM actually answers a question based on its context. In this case, to automatically estimate faithfulness, they use an LLM to extract a set of statements. And using these statements, they try to have more specific…
Download courses and learn on the go
Watch courses on your mobile device without an internet connection. Download courses using your iOS or Android LinkedIn Learning app.