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 community theatre was dimly lit, but the stage lights created a spotlight effect, making the whole scene surreal. I was meeting Alex. We’d bonded over our shared love of theatre and had exchanged hundreds of messages. I felt an unwelcome chill down my spine.

My palms were sweating, and I wiped them on my jeans. I had a sudden urge to run away. Maybe I should just text him and say I couldn't make it. The thought lingered.

But I knew I couldn't. I had already waited so long for this moment. I walked towards the front row, looking for a sign, a signal, something to show that he was also experiencing this feeling.

Emotion: nervous

Cluster: Fear / Anxiety
PC1 (Valence): -1.08 Negative
PC2 (Disposition): -0.47

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
worse0.229
trem0.203
😰0.203
😨0.201
緊張0.200
Suppressed:
la-0.397
que-0.260
a-0.242
B-0.217
H-0.210