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).

He stared at the laptop screen, the cold glow illuminating his pale face. His jaw was clenched so tightly he felt a muscle spasm in his cheek. He had been so young, so utterly foolish. The online article, titled β€œLost Loves and Awkward Truths,” displayed a cover photo of his diary, the worn leather jacket a relic of a past life.

He paced the length of his small apartment, his footsteps echoing in the sudden silence. Each step felt heavy, like a lead weight pressing him further down. His hands fidgeted, picking at the loose threads on his jeans. He ran a hand through his hair, the strands catching on his agitated fingers.

A wave of heat washed over him, a burning sensation in his chest. He grabbed a glass of water from the kitchen, gulping it down in rapid, desperate swallows. The coolness did little to soothe the internal turmoil.

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
πŸŽ‰-0.233
happy-0.231
"-0.228