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 scent of baking cookies filled the house, a smell that usually brought him comfort. Today, it only deepened the sense of something amiss. He was making cupcakes for his own party, a casual barbecue. Maya's party? A massive concert. He felt his jaw clench, a knot forming in his stomach.

He moved about the kitchen slowly, his actions mechanical. He measured ingredients with a perfunctory air. The bright colors of the sprinkles seemed to mock him. He felt as if he was watching himself from a distance, a detached observer of his own life.

He sampled the batter, tasting nothing. The sweetness, which usually brought him joy, was bland and uninteresting. He just wanted the day to be over.

Emotion: gloomy

Cluster: Sadness / Despair
PC1 (Valence): -1.68 Negative
PC2 (Disposition): -1.37

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
😞0.368
无法0.319
😔0.300
いつも0.270
쓸0.255
Suppressed:
la-0.432
l-0.419
de-0.398
B-0.294
!-0.259