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 gallery's white walls swam before Elara’s eyes, the starkness amplifying the tremor in her hands. She clutched her handbag so tightly her knuckles whitened. The vibrant canvases, *her* canvases, stared back at her with a disconcerting familiarity. She hadn't given permission.

A bead of sweat trickled down her temple. The crowd milled about, murmuring praises she hadn’t earned, accolades for a talent she hadn’t shared. She hadn't even finished the damn series.

"Extraordinary color palettes!" a woman gushed, pointing at a painting of a weeping willow. Elara's heart hammered against her ribs, a frantic bird desperate for escape. She wanted to scream, to yank them from the walls, to disappear.

Emotion: unnerved

Cluster: Fear / Anxiety
PC1 (Valence): -2.36 Negative
PC2 (Disposition): -0.55

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
😰0.302
😣0.282
будто0.265
😖0.265
ness0.257
Suppressed:
de-0.585
la-0.495
a-0.403
happy-0.289
ecstatic-0.284