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 color drained from Amelia’s face. She stared at the website, the words blurring before her eyes. The glossy photos and impressive testimonials were a far cry from the reality she'd just witnessed. Her friend, Chloe, had always bragged about her cutting-edge marketing company. But that "company"? It was a tiny, dilapidated office above a laundromat.

Amelia felt a cold sweat prickle her skin. The bravado, the endless stories of high-profile clients… it was all a fabrication. Now, she felt a profound sense of… what? Betrayal? Anger? Disappointment?

She slammed her laptop shut. The sudden movement released a small whimper from her. She needed to be alone, to process this. She grabbed her coat and fled from her apartment, the silence suddenly deafening.

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