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).

“It’s simply not possible,” Martha repeated, her voice thin, almost a whisper. The genetic testing results, the ones that indicated she carried the BRCA gene, the ones that had led her to consider a mastectomy… were incorrect. They belonged to someone else.

She felt a wave of dizziness, a sensation of being pulled under. The relief was a cold and distant thing, shadowed by a different terror. She had made so many decisions, consulted with so many doctors, shared so many emotions, based on what now felt like a cruel joke.

She buried her face in her hands, her shoulders shaking. The room swam around her, the clinical white of the doctor’s office blurring into a hazy mess.

Emotion: vulnerable

Cluster: Fear / Anxiety
PC1 (Valence): -2.07 Negative
PC2 (Disposition): -0.69

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
S0.327
ness0.319
😞0.299
😔0.277
😣0.258
Suppressed:
de-0.567
la-0.488
l-0.271
(!)-0.267
(!)-0.257