Project Overview


Through our experiments, we show that our 7B parameter ClimateGPT model outperforms, on climate tasks, a state of the art 70B model. These results demonstrate that we can reuse and adapt a high performing foundational model to a specific domain, such as climate, efficiently while also maintaining performance on standard tasks.

Overall Climate Task Performance


Cookie Settings

We use cookies to provide you with the best possible experience. They also allow us to analyze user behavior in order to constantly improve the website for you.