LLMs for Social Science

From embeddings
to agents.

Language models are reshaping how social scientists collect, annotate, and analyze text. This course builds the conceptual depth and hands-on skills to use them critically. Beginner Python required; everything else is taught from first principles.

§ 01 · Modules

Five weeks of material

What you will learn.

The course runs from mathematical foundations to autonomous research agents. Each module builds on the last; start at Module 1.

Module 1

Foundations

From bag-of-words to embeddings and tokenization, then contextual representations, language modeling, the Transformer, scaling laws, and how this stack shows up in social science workflows. The conceptual spine for everything that follows.

Open module →

Module 2

From Models to Tools

Post-training alignment (SFT, RLHF, DPO), prompting from zero-shot to chain-of-thought, reasoning techniques, and how to evaluate models and choose one that fits your task.

Open module →

Module 3

Deploying for Research

Fine-tuning (including LoRA-style adaptation and encoder heads), moving from notebooks to APIs and self-hosted inference, and rigorous validation — so a classifier works as a measurement instrument you can defend in publication.

Open module →

Module 4

Extraction, Summarization, and RAG

Faithful extraction and summarization (hallucination, omission, and distortion), then retrieval-augmented generation: chunking, embeddings, retrieval, and grounded answers with provenance checks over your own corpora.

Open module →

Module 5

Agentic Workflows

Building autonomous research agents with tool use, ReAct patterns, and multi-step orchestration. The frontier of what language models can do for your research pipeline.

Open module →

§ 02 · Begin

Getting started

Open access. No registration required.

Prerequisites

Beginner-to-intermediate Python. A Google account for Colab. No prior deep learning or NLP experience required; the course builds from mathematical foundations.

Preliminary notebooks →