Toy-model experiments on how training configurations affect feature-entanglement proxies in neural representations.
-
Updated
Sep 10, 2026 - Python
#
polysemanticity
Here are 2 public repositories matching this topic...
Reproducible causal-subspace experiments in GPT-2-small: interventions, matched-span controls, and audit-ready evidence.
pytorch computational-neuroscience superposition causal-inference ai-safety interpretability gpt-2 toy-models neuroai mechanistic-interpretability transformer-circuits polysemanticity mixed-selectivity
-
Updated
Aug 14, 2026 - Python
Add this topic to your repo
To associate your repository with the polysemanticity topic, visit your repo's landing page and select "manage topics."