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 glossy magazine, "The Admissions Insider," lay on Emily’s coffee table. She flipped through it, idly, until she saw it - her essay. It was titled, “Overcoming Obstacles,” and her heart sank.

She felt a tightening in her chest. She had poured her heart into that essay, the story of her family's struggles. Now, her most private grief was being presented as a success story, a testament to her resilience.

She felt a wave of dizziness. She gripped the edge of the coffee table, needing to steady herself. The feeling was a tight knot in the pit of her stomach. She felt the blood rush to her face.

Emily wanted to scream, to tear the magazine apart. The feeling was a sick, sinking feeling. She pictured her relatives reading it, analyzing her life. The pain of the past was now a public display. She needed to leave the room.

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