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).

He found himself pacing the worn rug in his inherited study. The news had dropped like a lead weight in his stomach. His father, the man he'd thought he knew, had kept a secret his entire life. A secret that now, according to the letter clutched in his sweaty hand, included a half-brother.

The room seemed to shrink. He ran a hand through his hair, the gesture more a reflex of distress than a conscious action. He glanced at the family portraits on the wall, and the faces of his ancestors seemed to mock him. A sense of displacement washed over him, a feeling like he no longer truly belonged in this space, in this lineage.

He went to the liquor cabinet, though he wasn’t usually a drinker. The amber liquid glinted invitingly. He poured a generous amount into a tumbler and gulped it down, the burn doing nothing to quell the rising tide of internal turmoil.

Emotion: unsettled

Cluster: Fear / Anxiety
PC1 (Valence): -2.33 Negative
PC2 (Disposition): -0.65

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
S0.382
😞0.309
ness0.301
😣0.256
😔0.248
Suppressed:
de-0.721
la-0.597
a-0.365
"-0.299
l-0.274