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).

"You're… you're saying I'm not who I thought I was?" Sarah stammered, her voice barely a whisper. The genetic counselor across the table nodded, her expression sympathetic but professional. The confirmation had been delivered with practiced tact, but for Sarah, it felt like a punch to the face.

She gripped the edge of the table, her knuckles white. The sterile environment of the office seemed to amplify the unsettling silence. The foundation of her life, the bedrock of her identity, had crumbled in an instant. She felt like she was teetering on the edge of a great cliff. Her body felt cold.

Emotion: rattled

Cluster: Fear / Anxiety
PC1 (Valence): -2.29 Negative
PC2 (Disposition): -0.42

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
S0.264
😣0.261
worse0.256
😞0.248
πŸ˜–0.247
Suppressed:
de-0.571
la-0.414
a-0.377
happy-0.259
enthusi-0.234