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).

He’d chosen the worst possible place. The announcement, delivered with a forced cheerfulness, landed like a punch in the gut. They were at the rooftop bar, overlooking the city they’d both sworn to conquer together. Now, he was going to abandon ship.

“Think of all the new experiences!” Mark said, grinning. A hollow laugh escaped her. The city lights blurred, reflecting in the sheen of the cocktail she’d barely touched. Every gaze in the crowded bar seemed to be locked on her.

Her palms began to sweat, clinging to the cold glass. She found herself fiddling with the strap of her purse, a nervous habit she hadn’t indulged in years. The music felt too loud, the laughter abrasive. It felt like she was the only one in the room not celebrating.

Emotion: humiliated

Cluster: Embarrassment
PC1 (Valence): -1.79 Negative
PC2 (Disposition): 0.93

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
C0.370
l0.329
尴尬0.322
ness0.303
because0.298
Suppressed:
de-0.369
own-0.254
😍-0.253
🤩-0.245
cautiously-0.228