Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

LEAF: Learning and Evaluation Augmented by Fact-Checking to Improve Factualness in Large Language Models

Hieu Tran, Junda Wang, Yujan Ting, Weijing Huang, Terrence Chen

TL;DR

It is suggested that integrating fact-checked responses whether through RAG enhancement or self-training enhances the reliability and factual correctness of LLM outputs, offering a promising solution for applications where information accuracy is crucial.

Abstract

Large language models (LLMs) have shown remarkable capabilities in various natural language processing tasks, yet they often struggle with maintaining factual accuracy, particularly in knowledge-intensive domains like healthcare. This study introduces LEAF: Learning and Evaluation Augmented by Fact-Checking, a novel approach designed to enhance the factual reliability of LLMs, with a focus on medical question answering (QA). LEAF utilizes a dual strategy to enhance the factual accuracy of responses from models such as Llama 3 70B Instruct and Llama 3 8B Instruct. The first strategy, Fact-Check-Then-RAG, improves Retrieval-Augmented Generation (RAG) by incorporating fact-checking results to guide the retrieval process without updating model parameters. The second strategy, Learning from Fact-Checks via Self-Training, involves supervised fine-tuning (SFT) on fact-checked responses or applying Simple Preference Optimization (SimPO) with fact-checking as a ranking mechanism, both updating LLM parameters from supervision. These findings suggest that integrating fact-checked responses whether through RAG enhancement or self-training enhances the reliability and factual correctness of LLM outputs, offering a promising solution for applications where information accuracy is crucial.

LEAF: Learning and Evaluation Augmented by Fact-Checking to Improve Factualness in Large Language Models

TL;DR

It is suggested that integrating fact-checked responses whether through RAG enhancement or self-training enhances the reliability and factual correctness of LLM outputs, offering a promising solution for applications where information accuracy is crucial.

Abstract

Large language models (LLMs) have shown remarkable capabilities in various natural language processing tasks, yet they often struggle with maintaining factual accuracy, particularly in knowledge-intensive domains like healthcare. This study introduces LEAF: Learning and Evaluation Augmented by Fact-Checking, a novel approach designed to enhance the factual reliability of LLMs, with a focus on medical question answering (QA). LEAF utilizes a dual strategy to enhance the factual accuracy of responses from models such as Llama 3 70B Instruct and Llama 3 8B Instruct. The first strategy, Fact-Check-Then-RAG, improves Retrieval-Augmented Generation (RAG) by incorporating fact-checking results to guide the retrieval process without updating model parameters. The second strategy, Learning from Fact-Checks via Self-Training, involves supervised fine-tuning (SFT) on fact-checked responses or applying Simple Preference Optimization (SimPO) with fact-checking as a ranking mechanism, both updating LLM parameters from supervision. These findings suggest that integrating fact-checked responses whether through RAG enhancement or self-training enhances the reliability and factual correctness of LLM outputs, offering a promising solution for applications where information accuracy is crucial.

Paper Structure

This paper contains 31 sections, 1 equation, 9 figures, 8 tables.

Table of Contents

  1. Introduction
  2. Methodology
  3. Fact-Checking for LLM Responses
  4. Calculation of Fact-Check Scores
  5. LEAF Mechanism I: Fact-Check-Then-RAG
  6. LEAF Mechanism II: Learning from Fact-Check via Self-Training
  7. Supervised Fine-Tuning on Factually Correct Responses
  8. Optimization with SimPO
  9. Experiments
  10. Filtered Accuracy via Fact-Checking
  11. Fact-Check-Then-RAG
  12. Supervised Fine-Tuning on Factually Correct Responses
  13. Fact-Checking as a Ranking Model
  14. SimPO on Ranked Responses
  15. Related Work
  16. ...and 16 more sections

Figures (9)

  • Figure 1: Comparison of workflows: standard LLM workflow (left), RAG-enhanced LLM workflow (middle), and our proposed Fact-Checking integrated workflow (right).
  • Figure 2: Fact-Check-Then-RAG is able to change the answer of LLMs by leveraging the knowledge retrieved from fact-check stage to regenerate the responses.
  • Figure 3: An Example Prompt for Query Generation with Context
  • Figure 4: An example query to MedRAG Corpus and 3 retrieved documents
  • Figure 5: An example prompt for Fact-Check with context. The final answer to the statement is [Not Supported].
  • ...and 4 more figures