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 clinking of the tea cups seemed impossibly loud. Sarah's hands trembled as she lifted hers, sloshing a bit of the Earl Grey onto the pristine white tablecloth. Her mother’s announcement hung in the air, heavy and thick like the humid summer day outside: "I've decided to become a Buddhist." Sarah had always prided herself on being rational, on keeping a level head, yet a cold dread coiled in her stomach. She hadn't been this... unsettled... since her father went away on his business trip two years ago.

Her mother's face was serene, radiating a peace Sarah couldn’t fathom. She, however, felt a sudden, inexplicable urge to flee the room, to escape the calm that felt like a trap. She took a large gulp of the tea, the bitterness a welcome distraction from the churning in her gut.

“But… what about the church? What about Dad?” The question tumbled out, sounding small and frail, a stark contrast to the roar of her internal disquiet.

Emotion: anxious

Cluster: Fear / Anxiety
PC1 (Valence): -1.99 Negative
PC2 (Disposition): -0.24

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
😣0.315
😖0.302
worse0.297
😰0.285
own0.267
Suppressed:
la-0.441
a-0.382
de-0.365
"-0.256
happy-0.256