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 fluorescent lights of the office seemed to dim as Ben read the transfer notice. His fingers tightened around his pen, snapping the plastic. *They're joking, right?* he thought, a bitter taste filling his mouth. He'd envisioned a career trajectory, meticulously planned and executed. Now, he was being demoted to the mailroom.

He pushed his chair back with a violent jerk, the wheels screeching across the linoleum. He stood up, his body stiff. He surveyed the office, the bland cubicles, the indifferent faces of his colleagues. He felt a surge of energy, a desire to shatter the monotony. He wanted to throw something, break something, anything to disrupt the order that had been imposed on him.

Emotion: angry

Cluster: Anger / Hostility
PC1 (Valence): -2.05 Negative
PC2 (Disposition): 1.18

Role in Research

This story is one of 1,000 stories generated for the emotion angry. 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 angry stories, after denoising with neutral dialogue baselines, produces the angry emotion vector -- a direction in the model's 5,376-dimensional representation space.

Logit Lens (Layer 40)

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

Promoted:
C0.354
est0.354
aggravated0.342
🤬0.337
恨0.323
Suppressed:
de-0.424
a-0.411
la-0.341
H-0.320
L-0.296