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).

He’d offered a half-hearted grin, the kind that didn't quite reach his eyes. "It's a great opportunity," Ben had said, his voice a little too loud, a little too eager. Mark had merely nodded, pushing a stray strand of hair from his forehead. The air in the brewery, usually filled with boisterous conversation and the clinking of glasses, felt suddenly thin and brittle.

His usual boisterous laughter was gone. The jokes he'd once shared with ease felt heavy, stuck in his throat. He ordered a pint of his favorite IPA, but barely tasted it. The familiar bitterness usually a comfort, now seemed like another reminder of the changes looming.

Later, walking home alone, the evening felt darker than it should. The streetlights cast elongated shadows, turning familiar landmarks into ominous shapes. He jammed his hands in his pockets, the cold seeping through his jeans. His usual route, a familiar path that once soothed him, felt alienating.

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