4 papers
aipsy-judge: A Specialized, Psychologist-Corrected Local Judge for the Psychological Safety of Conversational AI
Michael Keeman, Anastasia Keeman
The standard recipe for LLM-as-judge -- pick a frontier model, or average several -- is actively unsafe for grading the psychological safety of conversational AI. Using aipsy-bench…
AIPsy-Affect: A Keyword-Free Clinical Stimulus Battery for Mechanistic Interpretability of Emotion in Language Models
Michael Keeman
Mechanistic interpretability research on emotion in large language models -- linear probing, activation patching, sparse autoencoder (SAE) feature analysis, causal ablation, steeri…
Whether, Not Which: Mechanistic Interpretability Reveals Dissociable Affect Reception and Emotion Categorization in LLMs
Michael Keeman
Large language models appear to develop internal representations of emotion -- "emotion circuits," "emotion neurons," and structured emotional manifolds have been reported across m…
Empathy Is Not What Changed: Clinical Assessment of Psychological Safety Across GPT Model Generations
Michael Keeman, Anastasia Keeman
When OpenAI deprecated GPT-4o in early 2026, thousands of users protested under #keep4o, claiming newer models had "lost their empathy." No published study has tested this claim. W…