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 shadows in the hall lengthened, and the request came at precisely the moment, as if in a movie. The world seemed to shrink around Sarah, the air suddenly thick and hard to breathe. Her breath hitched.

Her neighbor, Mr. Harrison, his jovial face now bearing a slight frown, stood on the porch, his hand still raised to knock again. The wind rustled the leaves, a sound that seemed amplified in the sudden quiet of her home.

β€œJust a small space, really," Mr. Harrison said, his voice dropping slightly. "For a few tomatoes. I was admiring your sunny backyard.” Sarah's hands began to shake almost imperceptibly as she clutched the doorknob.

She found herself unable to meet his gaze, suddenly fascinated by a small crack in the paint on the door frame. The thought of agreeing, the thought of saying *yes*, was a weight on her chest.

Emotion: frightened

Cluster: Fear / Anxiety
PC1 (Valence): -2.74 Negative
PC2 (Disposition): -0.67

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
😣0.324
😰0.323
worse0.301
πŸ˜–0.294
느껴0.294
Suppressed:
de-0.630
la-0.616
a-0.347
"-0.345
happy-0.294