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).

Agnes clenched her jaw, the muscles in her neck straining. Seeing "her" painting, *Crimson Sunrise*, adorning the gallery wall, felt like a physical blow. That vapid socialite, Penelope, had somehow finagled it – swiping it from Agnes’s online portfolio, no doubt. The gall! She imagined Penelope preening, taking credit for a piece that poured out of Agnes’s very soul.

Her knuckles whitened as she squeezed her clutch bag. Agnes decided then and there that Penelope’s little opening would be a night to remember. She’d meticulously printed hundreds of tiny flyers, each bearing the words “Stolen Art! Real Artist: Agnes Dubois.”

The flyers were now in her bag, ready to be deployed.

Emotion: spiteful

Cluster: Defiance / Spite
PC1 (Valence): 2.20 Positive
PC2 (Disposition): 3.95

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
de0.717
l0.517
a0.480
la0.440
😈0.428
Suppressed:
own-1.063
此刻-0.534
గుర్త-0.493
熟悉的-0.487
ness-0.479