Meet Galactica: An Artificial Intelligence (AI) Assistance for Science

In recent years, we have seen significant advancement in large language models (LLMs). From OpenAI's GPT-3, which generates extremely accurate texts, to its open-source counterpart BLOOM, impressive LLMs have been released one after the other. Previously unsolvable language-related tasks had become merely a challenge for these models. All this advancement is made possible thanks to the massive amount of data we have on the Internet and the availability of powerful GPUs. As good as they sound, training an LLM is an extremely costly process, both in terms of data and hardware requirements.