你想蒸馏的下一个员工，何必是同事。蒸馏任何人的思维方式——心智模型、决策启发式、表达DNA。Distill how anyone thinks.

Use Garry Tan's exact Claude Code setup: 23 opinionated tools that serve as CEO, Designer, Eng Manager, Release Manager, Doc Engineer, and QA

Auto-backup tool for AI agent workspaces — syncs files to Git with scheduled backups, web dashboard & Telegram notifications. Built with Go.

Official implementation of GDPO: Group reward-Decoupled Normalization Policy Optimization for Multi-reward RL Optimization

[ACL 2026 Main] Official Repo for Paper "Which Reasoning Trajectories Teach Students to Reason Better? A Simple Metric of Informative Alignment“

Multiplex Thinking: Reasoning via Token-wise Branch-and-Merge

[TPAMI 2026] Offical Repository of "AtomThink: Multimodal Slow Thinking with Atomic Step Reasoning"

Yuan3.0: Mixture-of-Experts (MoE) Language Model

Official implementation of Selective Entropy Regularization (SIREN), proposed by paper 'Rethinking Entropy Regularization in Large Reasoning Models'.

Code and Datasets for reviewing of "SynAdapt: Learning Adaptive Reasoning in Large Language Models via Synthetic Continuous Chain-of-Thought"

A Recipe for Building LLM Reasoners to Solve Complex Instructions

Pretraining and inference code for a large-scale depth-recurrent language model

g1: Using Llama-3.1 70b on Groq to create o1-like reasoning chains

Official repository for the paper Number Cookbook: Number Understanding of Language Models and How to Improve It.

[NeurIPS 2025] Thinkless: LLM Learns When to Think

Official Repository for "Continuous Chain of Thought Enables Parallel Exploration and Reasoning"

ACL'2025: SoftCoT: Soft Chain-of-Thought for Efficient Reasoning with LLMs. and preprint: SoftCoT++: Test-Time Scaling with Soft Chain-of-Thought Reasoning

Official implementation of the NeurIPS 2025 paper "Soft Thinking: Unlocking the Reasoning Potential of LLMs in Continuous Concept Space"

[EMNLP 2025] LightThinker: Thinking Step-by-Step Compression

Official repository for "CODI: Compressing Chain-of-Thought into Continuous Space via Self-Distillation"

Connector-Aware Compact CoT (Synthetic Method For Reasoning Data)

[ICLR 2026] An official implementation of "SIM-CoT: Supervised Implicit Chain-of-Thought"

[EMNLP 2025] TokenSkip: Controllable Chain-of-Thought Compression in LLMs

[EMNLP 2025] Verification Engineering for RL in Instruction Following

REverse-Engineered Reasoning for Open-Ended Generation