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 wind howled outside, mirroring the turmoil within Ethan. He’d stumbled upon a box of old photographs while cleaning out his recently deceased grandmother’s house. He'd always known his grandfather as a quiet, reserved man who rarely spoke of his past. The photos showed a different picture.

One showed his grandfather, a hardened face, in the uniform of a soldier, clutching a bottle and laughing with a group of his comrades. The man he thought he knew was absent. A sense of displacement settled over him, a feeling that nothing was quite what it seemed.

He found himself pacing the creaking floorboards of the empty house, unable to shake off the sense of unreality. The house, usually a place of comfort, had become a cage of conflicting emotions.

Emotion: troubled

Cluster: Fear / Anxiety
PC1 (Valence): -2.20 Negative
PC2 (Disposition): -0.81

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
😞0.343
S0.317
ness0.302
😔0.294
😢0.272
Suppressed:
de-0.553
la-0.511
l-0.307
/-0.260
(!)-0.236