yangzhch6

ReSocratic. OptiBench and ReSocratic Synthesis Method

DARS. The official implemention of "Depth-Breadth Synergy in RLVR: Unlocking LLM Reasoning Gains with Adaptive Exploration" (ICML 2026)

Mirror-Critique. The official implemention of "Critique to Verify: Accurate and Honest Test-Time Scaling with RL-Trained Verifiers"

InterMWP. dataset & code of "LogicSolver: Towards Interpretable Math Word Problem Solving with Logical Prompt-enhanced Learning" in Findings of EMNLP 2022

AlignedCoT. Implementation of our paper "Speak Like a Native: Prompting Large Language Models in a Native Style"

TreeRPO. The official implemention of "TreeRPO: Tree Relative Policy Optimization"

Self-implemented-SVM. Implemention of SVM without using sklearn

UnbiasedMWP. An unbiased dataset and an unbiased training strategy for Math Word Problem. Repo of "Unbiased Math Word Problems Benchmark for Mitigating Solving Bias"

Accordion-Thinking. Training LLM to fold the rasoning process

Prune-OPD. The official implemention of "Prune-OPD: Efficient and Reliable On-Policy Distillation for Long-Horizon Reasoning"

rlvr-baseline. implemention of grpo & dr.grpo using verl

yangzhch6.github.io. homepage

AI4Math-Track3. ICML 2024 AI for Math Workshop, Track 3, Automated Optimization Problem-Solving with Code

Multi-cycle-CPU. A Multi-cycle CPU based on MIPS instruction set

ML_Notes. 一些学习笔记+面试问题

Reversi_AI. Three version AI of Reversi

OS_Lab. A monolithic kernel operating system implemented in my sophomore year