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 small, neat handwriting filled the page. “See the Taj Mahal.” “Learn to dance tango.” “Break free.” Break free. The words hit him like a physical blow. James had a knot in his chest.

He slammed the drawer shut, the sound echoing in the silent house. He wanted to lash out, to scream, to accuse. He did none of it. He began to pace.

He needed to get away, to clear his head. He pulled on his coat and walked out the door. The crisp, cold air did nothing to soothe his agitation. He kept walking, the words of the list replaying in his mind.

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