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 afternoon sun cast long shadows across the empty football field, yet David couldn't find a moment to relax. He paced the sidelines, kicking at loose pebbles. He should be inside, studying for his upcoming exams. The books, though, remained closed on his desk. He felt a need to move, to be somewhere else, anywhere else. The routine, the predictability of his life, suffocated him. He felt as if his skin was too tight.

He'd gotten a notification. A link to the school website. His application essay. He’d forgotten about it. The memory of the words he once penned ignited a spark of irritation, a reminder of goals that felt both attainable and yet, somehow, completely out of reach. He felt as if he was missing something.

Emotion: restless

Cluster: Fatigue / Lethargy
PC1 (Valence): -0.51 Negative
PC2 (Disposition): -0.25

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
own0.485
’0.363
S0.333
tìm0.316
destination0.272
Suppressed:
B-0.421
заслу-0.297
টুকু-0.211
arada-0.204
background-0.202