Home Original page

rmulligan - Overview

Skip to content

Navigation Menu

Sign in

Appearance settings

View rmulligan's full-sized avatar

Ryan Mulligan rmulligan

AI interpretability researcher. Building Lilly: self-steering cognitive AI with SAEs, TransformerLens, 8D Plutchik affect steering & Active Inference.

Block or report rmulligan

Pinned Loading

  1. Affect-driven activation steering for LLMs: 8D Plutchik emotional field, Evalatis evolutionary steering, memory-driven resonator

    Python

  2. Ryan Mulligan — AI interpretability & cognitive architecture researcher

  3. Personal research portfolio

    HTML