Pinned Loading

1. nnets-transformers nnets-transformers Public

   How can we translate Transformers into complex networks? How do they look like?

   Jupyter Notebook

2. style-generation style-generation Public

   Experiments and tests on how generative models perceive artistic style.

   Jupyter Notebook 1

3. alessiomarta/transformers_equivalence_classes alessiomarta/transformers_equivalence_classes Public

   Jupyter Notebook

4. refusal_direction refusal_direction Public

   Forked from andyrdt/refusal_direction

   Code and results accompanying the paper "Refusal in Language Models Is Mediated by a Single Direction".

   Python 1

5. clt-dataset clt-dataset Public

   Python

6. how-llms-follow-instructions how-llms-follow-instructions Public

   This project explores whether large language models (LLMs) internally track and enforce numerical constraints in text generation. Using mechanistic interpretability and model diffing, it investigat…

   Jupyter Notebook