Dear friends,

Nice to meet you here! This is a toy demo of GLM-130B, an open bilingual pre-trained model from Tsinghua Univeristy. GLM-130B uses two different mask tokens: [MASK] for short blank filling and [gMASK] for left-to-right long text generation. When the input does not contain any MASK token, [gMASK] will be automatically appended to the end of the text. We recommend that you use [MASK] to try text fill-in-the-blank to reduce wait time (ideally within seconds without queuing).

This demo is a raw language model without instruction fine-tuning (which is applied to FLAN-* series) and RLHF (which is applied to ChatGPT); its ability is roughly between OpenAI davinci and text-davinci-001. Thus, it is currently worse than ChatGPT and other instruction fine-tuned models :(

However, we are sparing no effort to improve it, and its updated versions will meet you soon! If you find the open-source effort useful, please star our GitHub repo to encourage our following development :)

Generation Parameter

Disclaimer inspired from BLOOM

GLM-130B was trained on web-crawled data, so it's hard to predict how GLM-130B will respond to particular prompts; harmful or otherwise offensive content may occur without warning. We prohibit users from knowingly generating or allowing others to knowingly generate harmful content, including Hateful, Harassment, Violence, Adult, Political, Deception, etc.