Publications

PreFlect: From Retrospective to Prospective Reflection in Large Language Model Agents

Published in Preprint / Under Review, 2026

We develop a framework that shifts reflection from reactive, post-hoc correction to proactive, pre-execution foresight, anchoring a planning-phase reflection loop with a distilled taxonomy of historical errors.

Recommended citation: Hanyu Wang, Yuanpu Cao, Lu Lin, Jinghui Chen. (2025). "PreFlect: From Retrospective to Prospective Reflection in Large Language Model Agents." Under Review.

TruthFlow: Truthful LLM Generation via Representation Flow Correction

Published in Proceedings of the 42nd International Conference on Machine Learning (ICML 2025), 2025

We propose TruthFlow, a representation intervention framework that mitigates LLM hallucination by generating query-specific correction vectors via flow matching, enhanced by a truth-related subspace projection using SVD.

Recommended citation: Hanyu Wang, Bochuan Cao, Yuanpu Cao, Jinghui Chen. (2025). "TruthFlow: Truthful LLM Generation via Representation Flow Correction." ICML 2025. https://arxiv.org/abs/2502.04556

On the Discrepancy and Connection between Memorization and Generation in Diffusion Models

Published in ICML 2024 Workshop on Foundation Models in the Wild, 2024

We address the paradox in diffusion models where the theoretical optimal score suggests memorization while trained models generalize, attributing generalization to the superior geometric smoothness of the trained score function.

Recommended citation: Hanyu Wang, Yujin Han, Difan Zou. (2024). "On the Discrepancy and Connection between Memorization and Generation in Diffusion Models." ICML 2024 Workshop on Foundation Models in the Wild.