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 weight of the paper in David’s hand felt heavy, crushing. The account statement, detailing the sum his wife, Carol, had squirreled away, felt like a judgment. He'd never noticed the small, locked box in her closet, the one where the paperwork must have been kept.

He felt a sudden, inexplicable urge to flee. He grabbed his keys and jacket and headed for the door. Outside, the familiar street looked distorted, almost menacing. He felt a deep unease, the kind that made the world feel tilted at a disorienting angle.

He felt a pull towards the local bar. It wasn’t a usual hangout for him, but today it was the only place he could think of to maybe…make sense of it.

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