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).

Liam drummed his fingers on the worn tabletop, the rhythm a frantic counterpoint to the quiet hum of the cafe. His grandfather, usually a man of booming laughter, sat stiffly across from him, sipping his coffee. The old man had just shared the news: he was embracing Islam.

Liam found himself unable to meet his grandfather’s gaze. The cafe’s familiar sounds – the hiss of the espresso machine, the murmur of other conversations – now seemed distant, muffled. He fiddled with the sugar packets, peeling the paper with unnecessary force.

"Are you… happy, Grandpa?" The question felt like it was wrung from him. His grandfather smiled, a small, tight movement of his lips. Liam stared out the window, at the passing cars and the blurred cityscape, suddenly seeing the world differently, unsure of what the shift meant.

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