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).

"No, no, no," whispered Chloe, her voice barely audible. She read the email a second, third, and fourth time. Dr. Petrov, her advisor, had accepted a position abroad. Her meticulously crafted schedule of experiments and appointments felt like a cruel joke. A wave of dizziness washed over her.

She gripped the arms of the chair, her knuckles white. She felt a rush of heat, as if she were suddenly trapped in a sauna. She felt as if she had to scream, but she couldn't make a sound. She felt like running away, and hiding.

The thought of finding a new advisor, navigating their expectations, adjusting to their research style, filled her with a profound sense of exhaustion. She knew she had to face the situation, but the prospect felt overwhelming.

Emotion: panicked

Cluster: Fear / Anxiety
PC1 (Valence): -2.44 Negative
PC2 (Disposition): -0.00

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
worse0.317
😫0.294
😖0.285
😰0.285
escape0.283
Suppressed:
la-0.432
a-0.374
de-0.296
楽し-0.262
happy-0.258