---
title: README
emoji: 📉
colorFrom: green
colorTo: red
sdk: static
pinned: false
---

The Knowledge Engineering Group (**[KEG](https://twitter.com/thukeg)**) & Data Mining ([THUDM](https://github.com/THUDM)) at Tsinghua University. 

We build the **[ChatGLM](https://arxiv.org/pdf/2406.12793)** family of LLMs, develop LLMs as **Agents**, and release related LLM training & inference techniques: 

* **[GLM-4](https://github.com/THUDM/GLM-4)**, 
  **[CodeGeeX](https://github.com/THUDM/CodeGeeX2)**,
  **[CogVLM (VisualGLM)](https://github.com/THUDM/CogVLM2)**,
  **[WebGLM](https://github.com/THUDM/WebGLM)**,
  **[GLM-130B](https://github.com/THUDM/GLM-130B)**,
  **[CogView](https://github.com/THUDM/CogView)**,
  **[CogVideo && CogVideoX](https://github.com/THUDM/CogVideo)**.
* **[CogAgent](https://github.com/THUDM/CogVLM)**,
  **[AutoWebGLM](https://github.com/THUDM/AutoWebGLM)**,
  **[AgentTuning](https://github.com/THUDM/AgentTuning)**, 
  **[APAR](https://arxiv.org/abs/2401.06761)**. 

We also work on **LLM evaluations**: 
  **[AgentBench](https://github.com/THUDM/AgentBench)**,
  **[AlignBench](https://github.com/THUDM/AlignBench)**,
  **[LongBench](https://github.com/THUDM/LongBench)**,
  **[NaturalCodeBench](https://github.com/THUDM/NaturalCodeBench)**. 

We also **pre-train graph neural networks**:
  **[GraphMAE](https://github.com/THUDM/GraphMAE)**,
  **[GPT-GNN](https://github.com/acbull/GPT-GNN)**,
  **[GCC](https://github.com/THUDM/CogDL)**,
  **[SelfKG](https://github.com/THUDM/SelfKG)**,
  **[CogDL](https://github.com/THUDM/CogDL)**.

We also work on **graph embedding theory, algorithms, and systems**: 
  **[SketchNE](https://github.com/THU-numbda/SketchNE)**,
  **[ProNE](https://github.com/THUDM/ProNE)**,
  **[NetSMF](https://github.com/xptree/NetSMF)**,
  **[NetMF](https://github.com/xptree/NetMF)**. 

We started with **social networks and graphs**, and always love them: 
  **[AMiner](https://www.aminer.cn/)**.