AlignmentResearch/tuned-lens
Tools for understanding how transformer predictions are built layer-by-layer
Tuned Lens helps machine learning researchers understand how large language models make predictions. It takes an existing transformer model and reveals what the model 'thinks' at each internal step, even before it outputs the final word. This allows researchers to analyze the step-by-step reasoning process within the model.
574 stars. No commits in the last 6 months.
Use this if you want to deeply analyze the internal workings of transformer models to understand how they arrive at their final predictions.
Not ideal if you are looking to improve the performance of a model or train new models, as this tool focuses on interpretability rather than development.
Stars
574
Forks
62
Language
Python
License
MIT
Category
Last pushed
Aug 07, 2025
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/transformers/AlignmentResearch/tuned-lens"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
lucidrains/x-transformers
A concise but complete full-attention transformer with a set of promising experimental features...
kanishkamisra/minicons
Utility for behavioral and representational analyses of Language Models
lucidrains/simple-hierarchical-transformer
Experiments around a simple idea for inducing multiple hierarchical predictive model within a GPT
lucidrains/dreamer4
Implementation of Danijar's latest iteration for his Dreamer line of work
Nicolepcx/Transformers-in-Action
This is the corresponding code for the book Transformers in Action