Agent-operable mechanistic-interpretability lab on the Google Colab CLI — one-command paper replication, GPU research loops, and an agent skill file
By chatting or signing in you agree to the Terms and chat-message logging (revocable in History).