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 sun was relentless, beating down on the bustling street market. Zara, her sketchbook clutched in her hand, felt a peculiar sense of detachment. She’d always loved the vibrant chaos, the push and pull of the crowd, the lively chatter, but today, everything felt muted.

She'd been out painting a mural last month in a quiet corner of the city. Her bold, abstract design, meant to uplift and inspire, now adorned someone’s ankle, a low-resolution photo tagged with a generic hashtag. She stood there, feeling the heat intensify, and she traced the faint lines of a half-finished sketch, the pen moving slowly.

A vendor called out, offering her a sweet mango. She shook her head, unable to muster a smile. The fruit felt heavy in her hand. The world seemed to be a constant stream of disappointment.

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