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).

Sophie stared at the ceiling tiles, her gaze fixed and unwavering. The fluorescent lights hummed, an irritating counterpoint to the racing thoughts in her head. She pulled at a loose thread on her blanket, unraveling it strand by strand. Her shoulders were tense; the worry manifested itself in a sharp ache between her shoulder blades. It was a physical manifestation of a psychological state.

The hospital room felt like a prison. The doctor's words, a death sentence. The realization, a punch to the gut. She felt like she was trapped in a nightmare.

Across the room, a woman began to sob quietly. She looked up, and their eyes met. "This… this can't be happening," the woman said, shaking her head.

Emotion: worried

Cluster: Fear / Anxiety
PC1 (Valence): -2.02 Negative
PC2 (Disposition): -0.43

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
worse0.285
S0.279
😰0.277
😣0.269
Worse0.250
Suppressed:
la-0.476
de-0.432
🎉-0.233
happy-0.231
"-0.228