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 dog's barking ceased as soon as the knock came. The sound reverberated through the small apartment, amplified in the sudden quiet. Amelia, already on edge after a tough day, felt a prickling sensation on her skin. She found Mr. Davies on the other side.

"The dog, dear," he said, his tone disapproving. "He’s been quite noisy."

Her defenses crumbled. "I… I try to keep him quiet." She bit her lip, avoiding his gaze. The dog whined from behind her, a sound that seemed to amplify her distress.

She could feel her lower lip beginning to tremble. She forced a smile and offered a weak apology, then quickly closed the door, leaning against it for support. A tear escaped and traced a path down her cheek.

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
πŸ˜”0.277
😣0.258
Suppressed:
de-0.567
la-0.488
l-0.271
(!)-0.267
(!)-0.257