Main Purpose
Galactica is an open-source large language model trained on scientific knowledge with 120 billion parameters.
Key Features
- Trained on a large and curated corpus of scientific knowledge, including over 48 million papers, textbooks, and lecture notes.
- Can write whitepapers, reviews, Wikipedia pages, and code.
- Knows how to cite and write equations.
- Specialized on scientific knowledge, making it a valuable resource for science-related tasks.
- Offers prompt pre-training, which includes prompts extracted from the training datasets to maximize generality and performance on specific tasks.
Use Case
- Summarizing academic papers.
- Solving math problems.
- Generating Wiki articles.
- Writing scientific code.
- Annotating molecules and proteins.
