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 phone felt heavy in her hand. Emily stared at the screen, the cursor blinking, mocking her. Mark’s email, a casual dismissal of their childhood treehouse adventure, twisted her stomach. He wrote it off as a silly game, a childish fantasy. He downplayed the core of their shared experience.

She crumpled the email in her hand, the paper crackling like a furious whisper. The air in her small apartment seemed thick, suffocating. She threw the balled paper into the bin with a violent gesture. It was a complete lie.

She paced the room, a caged lioness. The memory, a precious jewel in her mind, had been cruelly shattered. She wouldn't let it go, wouldn't let him rewrite their history like it was nothing.

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