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 weight of the world seemed to be resting on David's shoulders. He stood in his kitchen, staring at the newspaper on the counter. There it was again, the digital version of the article his college essay, a model of excellence. He felt a strange disconnect, as if he was observing himself from a distance.

He didn't eat breakfast. His appetite had disappeared. He shuffled through his emails, avoiding the one from his mother asking how his classes were. He felt like he had been hollowed out.

He walked over to the back door, and his hands felt cold. He looked outside, the familiar sight of the garden offered no comfort. He slowly turned the knob and walked back inside, the thought of leaving his house filled him with dread.

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