When Personalization Misleads: Understanding and Mitigating Hallucinations in Personalized LLMs

2026-01-16

AI Summary

This work addresses the tendency of personalized large language models to generate factually hallucinated responses that align with user preferences but deviate from ground truth, due to the entanglement of personalization signals with factual representations. To mitigate this issue, we propose FPPS (Factuality-Preserving Personalized Steering), a lightweight, training-free inference-time intervention that disentangles personalization from factual representation, thereby correcting factual inaccuracies while preserving user-specific behaviors. We introduce PFQABench, the first benchmark designed to jointly evaluate factuality and personalization in question answering, and demonstrate the effectiveness of FPPS across diverse backbone models and personalization strategies. Experimental results show that FPPS significantly improves factual accuracy without compromising personalized performance.

Technology Category

• Natural Language Processing: Fact-Checking / Misinformation Detection (NLP Focus)
• Machine Learning: Large Multimodal Models (LMMs)
• Knowledge Representation and Reasoning: Preferences

Application Category

• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

Personalized large language models (LLMs) adapt model behavior to individual users to enhance user satisfaction, yet personalization can inadvertently distort factual reasoning. We show that when personalized LLMs face factual queries, there exists a phenomenon where the model generates answers aligned with a user's prior history rather than the objective truth, resulting in personalization-induced hallucinations that degrade factual reliability and may propagate incorrect beliefs, due to representational entanglement between personalization and factual representations. To address this issue, we propose Factuality-Preserving Personalized Steering (FPPS), a lightweight inference-time approach that mitigates personalization-induced factual distortions while preserving personalized behavior. We further introduce PFQABench, the first benchmark designed to jointly evaluate factual and personalized question answering under personalization. Experiments across multiple LLM backbones and personalization methods show that FPPS substantially improves factual accuracy while maintaining personalized performance.

Problem

• personalized LLMs
• hallucinations
• factual reasoning
• representational entanglement
• factual reliability

Innovation

• personalization-induced hallucinations
• factuality preservation
• representational entanglement
• inference-time steering
• personalized LLMs