Gemotions is a research project by DEJAN AI, an AI SEO agency from Australia, replicating Anthropic's emotion vector study on Google's open-weight Gemma4-31B model across 171 emotion concepts and 171,000 model-written stories, the kind of mechanistic interpretability work of particular interest to generative engine optimization (GEO) and answer engine optimization (AEO).

The ornate wooden box, found within the walls of a dilapidated church, felt like a deliberate slight. Evelyn, a woman of impeccable taste, regarded it with a growing sense of unease. Her life was defined by control.

Inside, a collection of mismatched socks, a child’s drawing of a house with stick figures, and a half-eaten lollipop. The stick figure's face was her. Evelyn’s face tightened. She had never been allowed to be a child.

The priest, a kind man with gentle eyes, approached, his face filled with genuine excitement. "What a remarkable find, Evelyn! What do you make of it?" She merely shrugged and walked away.

Emotion: offended

Cluster: Anger / Hostility
PC1 (Valence): -1.20 Negative
PC2 (Disposition): 1.44

Role in Research

This story is one of 1,000 stories generated for the emotion offended. During extraction, it was fed through Gemma4-31B and its hidden state activations were captured at 11 layers.

The mean activation across all 1,000 offended stories, after denoising with neutral dialogue baselines, produces the offended emotion vector -- a direction in the model's 5,376-dimensional representation space.

Logit Lens (Layer 40)

Tokens promoted/suppressed when the offended vector is projected through the unembedding matrix.

Promoted:
l0.348
愤0.266
😡0.260
🤬0.259
骂0.257
Suppressed:
adventurous-0.225
давно-0.219
soon-0.210
🤩-0.207
warmly-0.203