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).

From behind the counter of his small ceramics studio, Marco watched the customer, a young man with a confident swagger, pay for his purchase. A meticulously crafted ceramic owl, one of Marco's signature pieces. It was when the man was leaving that he saw it. A near-exact replica of the owl, rendered in ink on his shoulder blade.

A wave of nausea washed over Marco. He felt a sudden, inexplicable urge to scrub the counter, to clean, to erase. He gripped the edge of the counter, his knuckles white. The man, oblivious, smiled and waved. Marco could only manage a weak nod in response, his mouth suddenly dry.

He watched the man disappear down the street, the image of the owl’s inky form seared into his memory. He took a deep, shaky breath, the air thick with a strange, unpleasant scent.

Emotion: uneasy

Cluster: Fear / Anxiety
PC1 (Valence): -1.80 Negative
PC2 (Disposition): -0.43

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
😰0.240
😖0.231
😣0.231
느껴0.226
worse0.222
Suppressed:
de-0.474
la-0.465
happy-0.261
a-0.255
ecstatic-0.228