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 shovel bit deep into the earth, yielding to the pressure as Alex unearthed the metal box. He was restoring the garden of the old Victorian house, a job that should have been simple, but he felt a strange sense of disquiet. The box was small, unassuming, yet held the potential to reveal some forgotten truth. Inside, he found a sealed envelope, addressed in his own youthful scrawl.

The letter inside was filled with ambitious plans, declarations of love for a woman he could barely recall, and a fierce, unwavering belief in their future. The closing words, "Never give up on us," echoed in his mind. The letter felt alien, a stranger’s plea. Alex crumpled the paper, his knuckles white. The air felt thin, difficult to breathe. He wanted to smash something.

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