Posts by Collection

portfolio

publications

Causality-Enhanced Behavior Sequence Modeling in LLMs for Personalized Recommendation

Published in Under Review, 2024

Proposes Counterfactual Fine-Tuning (CFT), a fine-tuning method that leverages counterfactual information to better model user behavior sequences for personalized recommendation.

Recommended citation: Zhang, Y., You, J., Bai, Y., Zhang, J., Bao, K., Wang, W., Chua, T.-S. (2024). "Causality-Enhanced Behavior Sequence Modeling in LLMs for Personalized Recommendation." Under Review.
Download Paper

How Are Your Preferences Encoded? Neuron-level Mechanisms of LLM Personalization

Published in Under Review, 2025

This paper introduces the concept of preference neurons that encode preferences in LLM, providing detailed analysis of LLM behaviors under single and multiple preferences.

Recommended citation: You, J., Zhao, X., Zhang, Y., Li, M., Chen, Y., Feng, F., Cheng, H., He, X., Chua, T.-S. (2025). "How Are Your Preferences Encoded? Neuron-level Mechanisms of LLM Personalization." Under Review.

NextQuill: Causal Preference Modeling for Enhancing LLM Personalization

Published in International Conference on Learning Representations (ICLR), 2026

Proposes NextQuill, a novel causal preference modeling-based alignment method for LLM personalization.

Recommended citation: Zhao, X., You, J., Zhang, Y., Wang, W., Cheng, H., Feng, F., Ng, S.-K., Chua, T.-S. (2026). "NextQuill: Causal Preference Modeling for Enhancing LLM Personalization." ICLR.
Download Paper

talks

teaching

Teaching experience 1

Undergraduate course, University 1, Department, 2014

This is a description of a teaching experience. You can use markdown like any other post.

Teaching experience 2

Workshop, University 1, Department, 2015

This is a description of a teaching experience. You can use markdown like any other post.