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).

John sat on the dusty floor, surrounded by boxes of his past. He picked up a crumpled science project, his elementary school handwriting a childish scrawl. He felt a familiar knot of discomfort forming in his stomach.

He was a tough guy now, a man who didn't back down. But these memories, these reminders of his earlier years…they made him feel strangely small. He felt a sudden, sharp ache.

He couldn’t look at it anymore. He hurled the project against the wall, the paper scattering across the floor. He wanted to rage, to break something. He took a deep, ragged breath, fighting back the unwelcome tears.

Emotion: sensitive

Cluster: Fear / Anxiety
PC1 (Valence): -2.10 Negative
PC2 (Disposition): -0.75

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
ness0.304
😞0.282
😥0.272
😣0.271
own0.264
Suppressed:
la-0.679
de-0.660
a-0.281
"-0.280
(!)-0.242