Siwei Wang

Benefits and Pitfalls of Reinforcement Learning for Language Model Planning: A Theoretical Perspective

Understanding and Enhancing the Planning Capability of Language Models via Multi-Token Prediction

Combinatorial Rising Bandits

PaT: Planning-after-Trial for Efficient Code Generation

Layer-Informed Fine-Tuning via Three-Stage Functional Segmentation of LLMs

Mechanism Design for LLM Fine-tuning with Multiple Reward Models

Offline Learning for Combinatorial Multi-armed Bandits