GLM-130B

GLM-130B puts a 130-billion-parameter bilingual language model in the open research stack THUDM built around the General Language Model (GLM) pre-training