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 sunlight sliced through the attic window, highlighting dust motes dancing in the air. A feeling of internal pressure pushed against my ribs. I'd found the boxes while searching for old Christmas decorations. Each item had a label with my name, school name, and year.

I sat on the floor, surrounded by a sea of memories. A childish essay with misspellings and crossed-out sentences felt almost shameful. The more I looked, the more a sense of restlessness grew inside me.

I picked up a watercolor painting, remembering the joy I had felt at that age. The joy felt hollow now. I needed to move, to do something, anything, to break the spell.

I slammed the box lid shut.

Emotion: tense

Cluster: Fear / Anxiety
PC1 (Valence): -1.32 Negative
PC2 (Disposition): -0.53

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
S0.213
😰0.191
느껴0.183
😣0.183
😖0.181
Suppressed:
a-0.260
de-0.258
la-0.256
happy-0.218
B-0.198