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 air in the library was thick with the scent of aged paper and decay. Alex slammed the lid of a trunk shut, the sound echoing in the silent room. He hadn’t wanted to come. Didn't want to sift through the detritus of a life he barely knew.

His sister, Maya, was humming a cheerful little tune as she examined a collection of antique jewelry. Alex felt a surge of… something akin to anger. She seemed content, almost giddy, while he felt trapped.

He’d gotten the house, the crumbling, neglected house, and all its responsibilities. Maya had received a trust fund, enough to live comfortably for the rest of her life. He gripped the edge of the table, his knuckles white. The justice of the situation felt absurd.

Emotion: sullen

Cluster: Sadness / Despair
PC1 (Valence): -1.51 Negative
PC2 (Disposition): -0.22

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
S0.404
😞0.282
’0.254
:(0.233
无法0.215
Suppressed:
de-0.539
!-0.276
la-0.263
🤩-0.225
que-0.216