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 first nail gun shot ripped through the morning calm, a metallic shriek that sliced through Amelia’s carefully constructed peace. She slammed her cereal bowl onto the counter, the milk splashing in a white arc. “Good grief,” she muttered, already envisioning months of this. Her knuckles were white against the cold ceramic.

She marched to the window, peering through the blinds. Mr. Henderson, a man whose lawn always looked suspiciously perfect, was supervising a gaggle of young men. They were tearing at his roof with the zeal of hungry wolves. Amelia felt a heat bloom in her chest, a slow burn that spread to her face.

Amelia’s eyes narrowed. This wasn't just a renovation; it was an act of war, a personal affront to her tranquility. She grabbed her phone, her fingers fumbling with the screen as she composed a strongly-worded text to her HOA president.

Emotion: irate

Cluster: Anger / Hostility
PC1 (Valence): -1.67 Negative
PC2 (Disposition): 1.45

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
C0.430
🤬0.350
aggravated0.343
骂0.340
恨0.331
Suppressed:
H-0.292
soon-0.289
☺️-0.289
optimistic-0.258
latter-0.253