# Causal Direct Preference Optimization for Distributionally Robust Generative Recommendation Source: [Causal Direct Preference Optimization for Distributionally Robust Generative Recommendation]() Collection index: [llms.txt]() Data updated: 2026-05-29T05:00:12.873Z Authors: Chu Zhao, Enneng Yang, Jianzhe Zhao, Guibing Guo Published: 2026-03-21T03:48:23.000Z ## Abstract Direct Preference Optimization (DPO) guides large language models (LLMs) to generate recommendations aligned with user historical behavior distributions by minimizing preference alignment loss. However, our systematic empirical research and theoretical analysis reveal that DPO tends to amplify spurious correlations caused by environmental confounders during the alignment process, significantly undermining the generalization capability of LLM-based generative recommendation methods in out of distribution (OOD) scenarios. To mitigate this issue, we propose CausalDPO, an extension of DPO that incorporates a causal invariance learning mechanism. This method introduces a backdoor adjustment strategy during the preference alignment phase to eliminate interference from environmental confounders, explicitly models the latent environmental distribution using a soft clustering approach, and enhances robust consistency across diverse environments through invariance constraints. Theoretical analysis demonstrates that CausalDPO can effectively capture users stable preference structures across multiple environments, thereby improving the OOD generalization performance of LLM-based recommendation models. We conduct extensive experiments under four representative distribution shift settings to validate the effectiveness of CausalDPO, achieving an average performance improvement of 17.17% across four evaluation metrics. ## Recorded benchmark results A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. - [Generative Recommendation on ML OOD 10M](): Generative Recommendation; ML OOD 10M; Hit Rate @10: 61. - [Generative Recommendation on Yelp OOD 2018](): Generative Recommendation; Yelp OOD 2018; H@10: 1.53. - [Generative Recommendation on Book-Crossing OOD](): Generative Recommendation; Book-Crossing OOD; Hit Rate @ 10: 64. - [Generative Recommendation on Book-Crossing (test)](): Generative Recommendation; Book-Crossing (test); H@10: 1.77. - [Generative Recommendation Popularity and Noisy Shifts](): Generative Recommendation; Generative Recommendation Popularity and Noisy Shifts; H@10: 0.53. ## Original sources and implementation [Original paper]() [PDF]()