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 crisp autumn air bit at Sarah’s exposed skin as she stood in the sunlit garden, holding the list. Her mother had always been a practical woman, but this list… it was a revelation. *Travel across Europe by train.* Sarah’s fingers, cold and numb, traced the words. The wind rustled through the leaves overhead, creating a nervous whisper in her ears.

The second item was manageable: *Learn to knit*. The next entry, though, had her vision blurring: *Go on a blind date*. Sarah felt a sudden pressure behind her eyes, the familiar weight of dread. The final entry, written with a shaky hand, caused her to feel weak in her knees: *Swim with dolphins*. She didn't want to know. She really, really didn't.

Emotion: terrified

Cluster: Fear / Anxiety
PC1 (Valence): -2.97 Negative
PC2 (Disposition): -0.62

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
😣0.357
worse0.350
无法0.334
😰0.333
sickening0.322
Suppressed:
la-0.676
a-0.392
de-0.371
"-0.344
happy-0.309