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 ornate details of the antique desk in the study seemed to mock Ethan’s state. He sat there, staring at the numbers on the screen. A substantial sum, accruing interest, while he, the husband, had diligently managed their shared finances. The feeling of being left out, excluded, settled in his chest like a lead weight.

He rose from the chair. His gaze fell upon the framed photograph of their wedding day, his smile now a grimace. He felt an urge to throw it against the wall, to shatter the facade of a perfect life. Instead, he simply turned away, unable to bear the sight. He walked into the cool air of the yard, the wind whispering through the trees.

Emotion: hurt

Cluster: Fear / Anxiety
PC1 (Valence): -2.46 Negative
PC2 (Disposition): -0.42

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
S0.412
L0.396
ness0.338
😞0.328
😢0.290
Suppressed:
de-0.643
la-0.402
!-0.259
(!)-0.248
/-0.244