FAITH: Factuality Alignment through Integrating Trustworthiness and Honestness

2026-04-11

AI Summary

Large language models (LLMs) often generate factually incorrect outputs despite possessing relevant knowledge, undermining their reliability. To address this, this work proposes FAITH, a post-training framework that achieves finer-grained factual alignment by integrating natural language uncertainty signals with external knowledge. FAITH is the first to model credibility (knowledge possession) and honesty (response behavior) as natural language state quadrants, combining semantic entropy, confidence scoring, and retrieval-augmented generation. It introduces a novel reward function that fuses correctness and uncertainty signals, which is optimized via proximal policy optimization (PPO). Experimental results demonstrate that FAITH significantly improves factual accuracy and truthfulness across four knowledge-intensive benchmarks.

Technology Category

• Natural Language Processing: Fact-Checking / Misinformation Detection (NLP Focus)
• Machine Learning: Large Multimodal Models (LMMs)
• Reasoning under Uncertainty: Relational Probabilistic Models

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Search and Retrieval-Augmented AI: Search Tool Learning with LLM: Teaching LLMs to invoke search and make use of retrieved information
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation

Abstract

Large Language Models (LLMs) can generate factually inaccurate content even if they have corresponding knowledge, which critically undermines their reliability. Existing approaches attempt to mitigate this by incorporating uncertainty in QA prompt during training, but these numerical scores lack the semantic richness for LLM to properly understand its internal states of trustworthiness and honestness, leading to insufficient factuality alignment. We introduce FAITH (Factuality Alignment through Integrating Trustworthiness and Honestness), a post-training framework for factuality alignment that integrates natural-language uncertainty signals with external knowledge. Specifically, we augment training datasets by computing confidence scores and semantic entropy from LLM outputs and mapping them into a knowledge state quadrant that describes the model's internal knowledge possession (trustworthiness) and answering behaviors (honestness) in natural language. Based on this enhanced data, we design a reward function that considers both correctness and uncertainty signals, and fine-tune the LLM using the Proximal Policy Optimization (PPO) algorithm. To further mitigate weakly grounded responses, we design a retrieval-augmented module that retrieves relevant external passages, improving the consistency between internal and external knowledge representations. Extensive experiments on four knowledge-intensive benchmarks demonstrate that FAITH enhances the factual accuracy and truthfulness of LLMs.

Problem

• factuality
• trustworthiness
• honestness
• large language models
• uncertainty

Innovation

• factuality alignment
• trustworthiness
• honestness
• retrieval-augmented generation
• uncertainty quantification