AIPsy-Affect: A Keyword-Free Clinical Stimulus Battery for Mechanistic Interpretability of Emotion in Language Models

arXiv:2604.23719v2 Announce Type: cross Abstract: Mechanistic interpretability research on emotion in large language models -- linear probing, activation patching, sparse autoencoder (SAE) feature analysis, causal ablation, steering vector extraction -- depends on stimuli that contain the words for the emotions they test. When a probe fires on "I am furious", it is unclear whether the model has detected anger or detected the word "furious". The two readings have very different consequences for e