THUDM

CodeGeeX2: A More Powerful Multilingual Code Generation Model

An optimized deep prompt tuning strategy comparable to fine-tuning across scales and tasks

An open-sourced end-to-end VLM-based GUI Agent

LongBench v2 and LongBench (ACL 25'&24')

A novel method to tune language models. Codes and datasets for paper ``GPT understands, too''.

LongCite: Enabling LLMs to Generate Fine-grained Citations in Long-context QA

Paper list of NLP for recommender systems