Siyan Zhao

OPSD. Python

OPSD. On-Policy Self-Distillation for LLMs

prepacking. The source code of our work "Prepacking: A Simple Method for Fast Prefilling and Increased Throughput in Large Language Models" [AISTATS 2025]

ICL_decision_boundary. official code for paper Probing the Decision Boundaries of In-context Learning in Large Language Models. https://arxiv.org/abs/2406.11233 [NeurIPS 2024]

decision-stacks. Implementation of Decision Stacks: Flexible RL via Modular Generative Models [NeurIPS 2023]

few-shot_Cifar-10-CNN. CNN

Few-shot-Cifar-100-5-classes. few-shot: 5 randomly chosen classes from cifar 100 class are used to train a convolutional neural network until convergence.