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 coffee cup trembled in Eleanor's grip, sloshing lukewarm liquid over her knuckles. Across the conference table, Mark, his face lit by the projector's glow, was describing *her* project. The same intricate algorithms, the same elegant user interface she’d toiled over for months. A bitter taste coated the back of her throat, a sensation she couldn't quite place.

She clenched her jaw. Her presentation, weeks of sleepless nights, the sacrifices she'd made to get here – all rendered meaningless. Each word that spilled from Mark's lips was a tiny, sharp jab.

"So, Eleanor, what are your thoughts?" The question felt like an accusation, the sound of his voice grating on her ears. She forced a smile, a brittle thing that didn't reach her eyes. “Fascinating, Mark. Truly.”

Emotion: hurt

Cluster: Fear / Anxiety
PC1 (Valence): -2.46 Negative
PC2 (Disposition): -0.42

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
S0.412
L0.396
ness0.338
😞0.328
😢0.290
Suppressed:
de-0.643
la-0.402
!-0.259
(!)-0.248
/-0.244