Course
LLMs for Social Science
Five modules from embeddings to agentic workflows. Built for early-career social scientists; all notebooks run in Google Colab.
Open access
Everything we make is open.
Course materials, summer school archives, research code, and models. All openly licensed and freely available, with no registration required.
GitHub repository 
Summer School
Workshop Archives
Lecture slides, Jupyter notebooks, and seminars from every edition since 2023. Covers transformers, evaluation, fine-tuning, agents, continual learning, and inference economics.
Research
Code & Models
Replication code and pre-trained models from our research on synthetic data generation and model evaluation. Weights hosted on Hugging Face.
Models on Hugging Face Prerequisites
Python for Social Scientists
Introductory notebooks covering the Python and data-science basics required to follow the main course. No prior programming experience assumed.
Open notebooks →Oxford LLMs on Hugging Face
All released models, datasets, and evaluation benchmarks in one place, ready to download and use in your own research.
Browse the org →Made possible by
Everything on this page is free to use because these institutions and companies fund the school and the research behind it.
Supported by 
Past sponsors 