analiza aiqbench
ConfRAG introduces a fine-tuning strategy that significantly reduces LLM hallucinations by training models to express uncertainty, thereby triggering retrieval-augmented generation only when necessary. This approach lowers computational costs by over 30% while maintaining high accuracy on factuality benchmarks through confidence calibration and atomic factual data.
ConfRAG: Confidence-Guided Retrieval-Augmenting Generation
arXiv:2506.07309v3 Announce Type: replace Abstract: Can Large Language Models (LLMs) be trained to avoid hallucinating factual statements, and can Retrieval-Augmented Generation (RAG) be triggered only when necessary to reduce retrieval and computation costs? In this work, we address both challenges simultaneously. We introduce ConfQA, a fine-tuning strategy that reduces hallucination rates from 20-40% to below 5% across multiple factuality benchmarks. The approach is simple: when the model answers correctly, it is trained to output the answer; otherwise, it is trained to respond with "I am unsure". Two design choices make this training effective: (1) a dampening prompt ("answer only if you are confident") that explicitly discourages overconfident hallucinations, and (2) training data drawn from atomic factual statements (e.g., knowledge graph attribute values), which calibrates model confidence and yields robust generalization across domains and question types. Building on ConfQA, we propose ConfRAG, a triggering strategy that invokes RAG only when the model responses with unsure. This framework achieves accuracy above 95% in ideal case while reducing unnecessary external retrievals by over 30%.
Abstract:Can Large Language Models (LLMs) be trained to avoid hallucinating factual statements, and can Retrieval-Augmented Generation (RAG) be triggered only when necessary to reduce retrieval and computation costs? In this work, we address both challenges simultaneously. We introduce ConfQA, a fine-tuning strategy that reduces hallucination rates from 20-40% to below 5% across multiple factuality benchmarks. The approach is simple: when the model answers correctly, it is trained to output the answer; otherwise, it is trained to respond with "I am unsure". Two design choices make this training effective: (1) a dampening prompt ("answer only if you are confident") that explicitly discourages overconfident hallucinations, and (2) training data drawn from atomic factual statements (e.g., knowledge graph attribute values), which calibrates model confidence and yields robust generalization across domains and question types. Building on ConfQA, we propose ConfRAG, a triggering strategy that invokes RAG only when the model responses with unsure. This framework achieves accuracy above 95% in ideal case while reducing unnecessary external retrievals by over 30%.
Computation and Language (cs.CL)
arXiv:2506.07309 [cs.CL]
(or arXiv:2506.07309v3 [cs.CL] for this version)
https://doi.org/10.48550/arXiv.2506.07309 Focus to learn more arXiv-issued DOI via DataCite
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 259 cuvinte
Subiecte conexe
HyCE-RAG: Hypergraph Chain-of-Evidence Retrieval-Augmented Generation for Explainable Multi-hop Question Answering
arXiv · 28 iul. 2026
Self-Augmenting Retrieval for Diffusion Language Models
arXiv · 6 iun. 2026
Introducing Mistral OCR 4
Mistral AI · 23 iun. 2026
Modernizing and scaling support operations with generative AI on AWS
AWS · 2 sept. 2026