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).

"Are you sure?" she asked, her voice barely audible. Her stomach churned. The waitress, Sarah, nodded sadly. "Yeah, I heard it from Mr. Henderson himself. Last day's next Sunday." The diner, Hank's, was where she always came after her therapy sessions. It was her reset button, a place where she felt at ease.

Her appetite vanished. She picked at the fries on her plate, the once-sizzling grease suddenly unappetizing. She felt a need to escape, to leave the scene of the announcement.

She pushed back from the table. "I, uh, I think I'll just be going." She left a generous tip, a small act of appreciation and a hurried farewell. She walked, head down, feeling the sting of unshed tears.

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
마다0.314
worse0.313
😰0.303
恐惧0.290
Suppressed:
de-0.691
la-0.562
a-0.402
"-0.316
brag-0.293