Viewing profile — hyluo

GLM-5.2 is a turning point for coding agents. It's the first model a business would actually pay to replace Claude Opus with. We gave GLM-5.2 the ability to control its own context…

Looks pretty cool! I love the automatic trigger functionality

- We build the Thread Inference Model (TIM) based on the transformer architecture, and its dedicated runtime TIMRUN. - TIM + TIMRUN = Intelligent workflow generation, context engin…

- We build the Thread Inference Model (TIM) based on the transformer architecture, and its dedicated runtime TIMRUN. - TIM + TIMRUN = Intelligent workflow generation, context engin…

thoughts?

The paper introduces improved performance by prompting LLMs with "natural language embedded programs (NLEP)". No task-specific prompt is needed. Paper: https://arxiv.org/abs/2309.1…

Thanks! will replace "chatgpt" with "gpt-3.5-turbo".

It is a problem, but the same training set can be used to train commercial models.

The model outputs ChatGPT by connecting to a search engine - it's expected because it is fed with search results. Without search, SAIL-7B does not outperform ChatGPT. The interesti…

Sry our backend gpu servers are being upgraded and it caused some problem for our job submission system. Our slurm is not fixed yet, but I've brought the model back online for now …