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 request echoed in her mind, a constant, irritating buzz. The woman from next door, always so cheerful, always so… presumptuous. To think she could just… take a portion of her yard. The thought was infuriating.

She stood at the window, arms crossed tightly, watching the woman walk back to her house. The afternoon sun seemed to mock her, casting long shadows across the manicured lawn. She hated the smile she was trying to give her.

Her jaw clenched, and she turned away from the window, wanting to break free of this stifling situation. She wanted to scream. She wanted to disappear. She needed to do something, anything, to alleviate the mounting pressure.

Emotion: enraged

Cluster: Anger / Hostility
PC1 (Valence): -2.14 Negative
PC2 (Disposition): 1.33

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
C0.555
est0.370
aggravated0.362
骂0.359
🤬0.345
Suppressed:
a-0.468
de-0.358
la-0.312
latter-0.287
☺️-0.285