Generating Grounded Responses to Counter Misinformation via Learning Efficient Fine-Grained Critiques

2025-06-06

AI Summary

Large language models (LLMs) often generate factually erroneous refutations, undermining the reliability of automated misinformation mitigation. Method: We propose MisMitiFact—a lightweight, fine-grained critique model trained via knowledge distillation from scalable fact-checking data. It precisely detects errors at numerical, entity, and topical levels and generates concise, actionable feedback to guide LLMs toward fact-anchored refutations. Crucially, it replaces computationally expensive LLM-based self-feedback with a dedicated, parameter-efficient critique module. Contribution/Results: Experiments show MisMitiFact achieves refutation accuracy comparable to LLM self-feedback while reducing model parameters by over 90% and increasing feedback throughput by approximately 5×. This enables cost-effective, large-scale deployment for real-world misinformation governance.

Technology Category

• Natural Language Processing: Fact-Checking / Misinformation Detection (NLP Focus)
• Machine Learning: Large Multimodal Models (LMMs)
• Application Domains: Misinformation & Fake News

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Fake news and misinformation poses a significant threat to society, making efficient mitigation essential. However, manual fact-checking is costly and lacks scalability. Large Language Models (LLMs) offer promise in automating counter-response generation to mitigate misinformation, but a critical challenge lies in their tendency to hallucinate non-factual information. Existing models mainly rely on LLM self-feedback to reduce hallucination, but this approach is computationally expensive. In this paper, we propose MisMitiFact, Misinformation Mitigation grounded in Facts, an efficient framework for generating fact-grounded counter-responses at scale. MisMitiFact generates simple critique feedback to refine LLM outputs, ensuring responses are grounded in evidence. We develop lightweight, fine-grained critique models trained on data sourced from readily available fact-checking sites to identify and correct errors in key elements such as numerals, entities, and topics in LLM generations. Experiments show that MisMitiFact generates counter-responses of comparable quality to LLMs' self-feedback while using significantly smaller critique models. Importantly, it achieves ~5x increase in feedback generation throughput, making it highly suitable for cost-effective, large-scale misinformation mitigation. Code and LLM prompt templates are at https://github.com/xxfwin/MisMitiFact.

Problem

• Automate counter-response generation to combat misinformation effectively
• Reduce LLM hallucination in misinformation mitigation with fact-grounded critiques
• Improve scalability and cost-efficiency in generating evidence-based counter-responses

Innovation

• Generates fact-grounded counter-responses efficiently
• Uses lightweight fine-grained critique models
• Increases feedback throughput by 5x