EMNLP Findings2025
Benchmarking and Improving LLM Robustness for Personalized Generation
Chimaobi Okite, Naihao Deng, Kiran Bodipati, Huaidian Hou, Joyce Chai, Rada Mihalcea
Abstract
Recent years have witnessed a growing interest in personalizing the responses of large language models (LLMs). While existing evaluations primarily focus on whether a response aligns with a user's preferences, we argue that factuality is an equally important yet often overlooked dimension. In the context of personalization, we define a model as robust if its responses are both factually accurate and align with the user preferences. To assess this, we introduce PERG, a scalable framework for evaluating robustness of LLMs in personalization, along with a new dataset, PERGData. We evaluate fourteen models from five different model families using different prompting methods. Our findings show that current LLMs struggle with robust personalization: even the strongest models (GPT-4.1, LLaMA3-70B) fails to maintain correctness in 5% of previously successful cases without personalization, while smaller models (e.g., 7B-scale) can fail more than 20% of the time. Further analysis reveals that robustness is significantly affected by the nature of the query and the type of user preference. To mitigate these failures, we propose Pref-Aligner, a two-stage approach that improves robustness by an average of 25% across models. Our work highlights critical gaps in current evaluation practices and introduces tools and metrics to support more reliable, user-aligned LLM deployments. We open-source our code and datasets at: https://github.com/MichiganNLP/ Benchmark_Improve_LLM_ Robustness_in_Personalization
Ask Lune
Know this paper in minutes, not an afternoon.
Lune can read every page of it and the top-tier papers that cite it, and gives you its claims, its evidence and its limits, each one cited to the paper it comes from. No skimming, no AI slop.
1 top-tier paper has cited it since. Know what they found before your reviewers do.
Lune can draw on
the full paper · the top-tier papers that cite it · related work
Free to start. No credit card required.
Lineage
Cited by top-tier papers1
Builds on18
- Measuring Massive Multitask Language UnderstandingDan Hendrycks, Collin Burns, Steven Basart, Andy Zou et al.ICLR 2021 · 7,905 citations
- TruthfulQA: Measuring How Models Mimic Human FalsehoodsStephanie Lin, Jacob Hilton, Owain EvansACL 2022 · 3,228 citations
- Large Language Models Cannot Self-Correct Reasoning YetJie Huang, Xinyun Chen, Swaroop Mishra, Huaixiu Steven Zheng et al.ICLR 2024 · 858 citations
- Jury Learning: Integrating Dissenting Voices into Machine Learning ModelsMitchell L. Gordon, Michelle S. Lam, Joon Sung Park, Kayur Patel et al.CHI 2022 · 134 citations
- Aligner: Efficient Alignment by Learning to CorrectJiaming Ji, Boyuan Chen, Hantao Lou, Donghai Hong et al.NeurIPS 2024 · 115 citations
Related papers
- Do LLMs Recognize Your Preferences? Evaluating Personalized Preference Following in LLMsSiyan Zhao, Mingyi Hong, Yang Liu, Devamanyu Hazarika et al.ICLR 2025
- Learning Personalized Alignment for Evaluating Open-ended Text GenerationDanqing Wang, Kevin Yang, Hanlin Zhu, Xiaomeng Yang et al.EMNLP 2024 · 2 citations
- From 1, 000, 000 Users to Every User: Scaling Up Personalized Preference for User-level AlignmentJia-Nan Li, Jian Guan, Songhao Wu, Wei Wu et al.ACL 2026
- When Personalization Misleads: Understanding and Mitigating Hallucinations in Personalized LLMsZhongxiang Sun, Yi Zhan, Chenglei Shen, Weijie Yu et al.ACL Findings 2026
- Exploring the Potential of LLMs as Personalized Assistants: Dataset, Evaluation, and AnalysisJisoo Mok, Ik-hwan Kim, Sangkwon Park, Sungroh YoonACL 2025 · 10 citations
