Pinned Loading

1. iatrogenic_effect iatrogenic_effect Public

   Mech Interp Experiments: iantrogenic effects Llama-3.1-8B/70B base vs instruct

   Python 1

2. attentional_hijacking attentional_hijacking Public

   This repo contains the six core experiments that demonstrate and characterize the mechanism of attentional hijacking, across Gemma 3 4B, 12B, and 27B (IT and PT variants).

   Python

3. researchRalph researchRalph Public

   Autonomous research using multi-agent swarm for experiments

   Python 1

4. ICML_experiments ICML_experiments Public

   Salience-weighted attentional hijacking: ablation experiments for ICML MechInterp Workshop

   HTML

5. softmaxExperiments softmaxExperiments Public

   Mechanistic interpretability: Truth Jailbreak attentional hijacking experiments on transformers

   Python

6. mindreader mindreader Public

   Fine-tuned classifiers for chain-of-thought deception detection - training code and weights

   Python