Pinned Loading

1. Brain-Transformer-Equivalence Brain-Transformer-Equivalence Public

   Do LLMs think like brains? We test GPT-2, BERT, Mistral, DeepSeek & Qwen+SAE against EEG data. Sparse features yield a 4.3× alignment jump. Working paper included.

   Jupyter Notebook

2. D-S-in-LLMs D-S-in-LLMs Public

   Jupyter Notebook

3. HOTCOvsHOTCO2 HOTCOvsHOTCO2 Public

   Jupyter Notebook

4. LEWM LEWM Public

   Python