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).

He felt the sting behind his eyes before he even stepped inside the dingy establishment. The bell above the door chimed, a brittle, unwelcome sound, and he winced. The pawn shop owner, a man whose face was a roadmap of disappointments, barely glanced up from his newspaper.

The recent argument with his sister had left him exposed, vulnerable. Every little thing felt like a blow. He'd come to the pawn shop on a long shot, a desperate attempt to find something, anything, that would remind him of the good times.

He'd been searching for hours, his fingers tracing the cold glass, when a flash of light caught his eye. It was nestled amongst a pile of gaudy costume jewelry. He picked it up. The delicate setting, the way the light danced on the stone. His grandmother’s ring. His heart began to hammer erratically.

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