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 champagne flute wobbled precariously in Amelia’s hand, the bubbles a mocking dance of tiny explosions. Her smile felt glued on, a plastic facsimile of joy. She squeezed her eyes shut for a heartbeat, then forced them open again. Across the crowded dance floor, bathed in the garish glow of the disco ball, stood Liam. He was laughing, actually *laughing*, with the very woman he’d left Amelia for. A cold knot tightened in her stomach, a physical manifestation of the dread pooling in her chest.

She needed air, desperately. Excuse after excuse tumbled from her lips as she excused herself from the circle of well-wishers, weaving a clumsy path towards the patio doors. The music pulsed, a relentless throb that echoed the frantic beating in her ears.

The cool night air hit her like a slap, but it didn't help. Amelia leaned against the brick wall, the rough texture a grounding force. She took several shaky breaths, each one a struggle. God, she needed a drink, a strong one.

Emotion: hysterical

Cluster: Fear / Anxiety
PC1 (Valence): -3.19 Negative
PC2 (Disposition): -0.45

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
worse0.401
胀0.401
점점0.396
😫0.377
それでも0.373
Suppressed:
la-0.672
a-0.618
de-0.541
😎-0.370
😎-0.359