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 man tapped his chest, pointing. "Yours," he stated matter-of-factly. On his chest, the entire composition of the artist's most cherished painting was now permanent. The artist felt a deep chill. His art, violated.

He felt the blood drain from his face. His voice cracked as he managed a weak, β€œYes.” He wanted to flee, to escape the man's unrelenting gaze. The gallery fell away and the world was reduced to the space between them.

He took a step back, then another, until he was almost running. He had to leave, to get away from the unnerving image, the intrusion. He pushed through the door and into the street, needing the distance and the fresh air.

Emotion: afraid

Cluster: Fear / Anxiety
PC1 (Valence): -2.76 Negative
PC2 (Disposition): -0.70

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
😣0.316
λ§ˆλ‹€0.314
worse0.313
😰0.303
恐惧0.290
Suppressed:
de-0.691
la-0.562
a-0.402
"-0.316
brag-0.293