Hanze Guo

Hanze Guo

Research on LLM alignment and LLM for recommendation.

Hanze Guo

I work on large language models, value alignment, causal reasoning, and explainable recommendation.

Selected Publications (CCF A Venues)

1. Counterfactual Reasoning for Steerable Pluralistic Value Alignment of Large Language Models

NeurIPS 2025 (CCF A, Poster)
Hanze Guo, Jing Yao, Xiao Zhou, Xiaoyuan Yi, Xing Xie
arXiv

We propose COUPLE, a counterfactual reasoning framework for steerable pluralistic value alignment of large language models, supporting fine-grained control over pluralistic human values via structural causal models and counterfactual sampling.

2. SoREX: Towards Self-Explainable Social Recommendation with Relevant Ego-Path Extraction

ACM Transactions on Information Systems (TOIS), 2025 (CCF A)
Hanze Guo, Yijun Ma, Xiao Zhou
ACM DL · Code

We develop SoREX, a self-explainable GNN-based social recommendation framework that extracts factor-specific ego-paths as human-understandable explanations while maintaining strong recommendation accuracy.