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 thud had echoed in the hollow space inside him. It had reverberated through his teeth, through his very bones. It was an intrusion, a violation of the fragile peace he'd worked so hard to maintain.

He stood frozen in the doorway, the scent of wood and freshly turned earth filling his nostrils. The neighbor, a woman usually brimming with good cheer, was standing by the ruin, her face a mask of disbelief.

He had always found her a little intimidating with her high energy. He didn't want to be around anyone right now. He instinctively took a step back, a silent retreat into the sanctuary of his home.

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