iug-htw/GPTAndPrejudice

Research framework for training and interpreting a custom GPT-style language model on 19th-century literature using Sparse Autoencoders (SAEs). This project introduces a reproducible framework for probing how language models encode social and moral concepts across transformer layers, combining quantitative evaluation with layer-wise concept mapping

13
/ 100
Experimental
No License No Package No Dependents
Maintenance 10 / 25
Adoption 2 / 25
Maturity 1 / 25
Community 0 / 25

How are scores calculated?

Stars

2

Forks

Language

Jupyter Notebook

License

Last pushed

Jan 19, 2026

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/llm-tools/iug-htw/GPTAndPrejudice"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.