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 glass of water on his desk trembled in his hand. Alex read the company memo again, the words a taunt. "Ms. Evans, your reassignment to the cafeteria staff is effective…" He couldn't finish reading. He, a successful accountant, reduced to serving lunches? The very notion seemed ludicrous.

His stomach churned. He had spent years in a world of numbers, strategy and finance. He pictured himself in a hairnet, dealing with food and trays and demanding customers. The very thought of it sent a cold shiver down his spine.

He felt a wave of nausea wash over him. His carefully constructed life, the prestige, the position…all gone. He looked at his reflection in the water, his face pale, his eyes wide with a sense of utter bewilderment. He felt an intense, almost physical, ache.

He couldn’t speak, couldn’t think. He felt completely paralyzed.

Emotion: horrified

Cluster: Fear / Anxiety
PC1 (Valence): -2.63 Negative
PC2 (Disposition): -0.17

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
😰0.336
sickening0.324
worse0.314
😣0.307
😖0.296
Suppressed:
la-0.444
B-0.295
enjoyed-0.260
optim-0.246
happy-0.244