{"paper":{"title":"Ragas: Automated Evaluation of Retrieval Augmented Generation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"Ragas supplies reference-free metrics that score context relevance, faithfulness to retrieved passages, and answer relevance in RAG pipelines.","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Jithin James, Luis Espinosa-Anke, Shahul Es, Steven Schockaert","submitted_at":"2023-09-26T19:23:54Z","abstract_excerpt":"We introduce Ragas (Retrieval Augmented Generation Assessment), a framework for reference-free evaluation of Retrieval Augmented Generation (RAG) pipelines. RAG systems are composed of a retrieval and an LLM based generation module, and provide LLMs with knowledge from a reference textual database, which enables them to act as a natural language layer between a user and textual databases, reducing the risk of hallucinations. Evaluating RAG architectures is, however, challenging because there are several dimensions to consider: the ability of the retrieval system to identify relevant and focuse"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"We put forward a suite of metrics which can be used to evaluate these different dimensions without having to rely on ground truth human annotations.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That LLM-based judgments of context relevance, faithfulness, and answer relevance will correlate sufficiently with human judgments across domains and models without additional calibration or validation data.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"Ragas supplies reference-free metrics for measuring context relevance, faithfulness to retrieved passages, and answer quality in RAG pipelines.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Ragas supplies reference-free metrics that score context relevance, faithfulness to retrieved passages, and answer relevance in RAG pipelines.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"e8604ffb2e1165e3985ff08277d4af73b8bd189af2184a600778006a75c32a97"},"source":{"id":"2309.15217","kind":"arxiv","version":2},"verdict":{"id":"7b8b1f3e-2760-402d-b137-0f8f35e51ae5","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-16T21:33:03.116002Z","strongest_claim":"We put forward a suite of metrics which can be used to evaluate these different dimensions without having to rely on ground truth human annotations.","one_line_summary":"Ragas supplies reference-free metrics for measuring context relevance, faithfulness to retrieved passages, and answer quality in RAG pipelines.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That LLM-based judgments of context relevance, faithfulness, and answer relevance will correlate sufficiently with human judgments across domains and models without additional calibration or validation data.","pith_extraction_headline":"Ragas supplies reference-free metrics that score context relevance, faithfulness to retrieved passages, and answer relevance in RAG pipelines."},"references":{"count":3,"sample":[{"doi":"","year":2022,"title":"Sparks of Artificial General Intelligence: Early experiments with GPT-4","work_id":"a23cfe92-7f7c-424b-98d4-b386a83002fb","ref_index":1,"cited_arxiv_id":"2303.12712","is_internal_anchor":true},{"doi":"","year":2020,"title":"Halueval: A large-scale hallucination evaluation benchmark for large language models","work_id":"2cf6bc2d-aed3-4a58-879e-daf0de687940","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2023,"title":"SelfCheckGPT: Zero-Resource Black-Box Hallucination Detection for Generative Large Language Models","work_id":"90efebf6-4006-4176-9bf1-21a876571751","ref_index":3,"cited_arxiv_id":"2303.08896","is_internal_anchor":true}],"resolved_work":3,"snapshot_sha256":"4e9839eed543571b87bdca218a259dada870540fa790eee7398efb59804fbf27","internal_anchors":2},"formal_canon":{"evidence_count":3,"snapshot_sha256":"a3f195451793100092dea034b102bece97ffd22a817852f739338ab1cfff9e1a"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"}