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).

“Another round?” Ben asked, his voice strained. He hadn’t touched his beer. The pub felt suddenly stifling. Daniel, the supposed rock star, was a cashier. He still looked the part, wearing ripped jeans and sporting a shaggy, dyed-black haircut, but Ben knew.

Ben had always idolized Daniel, reveling in the tales of sold-out concerts and screaming fans. He’d lived vicariously through Daniel’s supposed glamorous life. Now, the music, the lights, the very air of the pub seemed tainted.

He took a slow, deliberate sip from his drink. The bitterness of the beer did nothing to cleanse the sour taste in his mouth. He felt a profound sense of disappointment, a hollowness where admiration had once resided. He wondered if he could ever trust anyone again.

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