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 attic air felt thick, heavy with the weight of decades. Amelia stood frozen, holding a school notebook. On the cover, her mother's name was written in perfect calligraphy. Inside, neat rows of math problems, perfectly executed diagrams, and stories written with an effortless flair. A bitter taste coated Amelia's tongue.

She’d always struggled with math, always felt clumsy with words. Her mother, she realized, had been the opposite, a perfect student.

"So many awards," Amelia whispered to herself, tracing the gold foil sticker on the cover of the notebook. She felt a burning sensation behind her eyes.

That night, she avoided her mother at dinner.

Emotion: envious

Cluster: Shame / Guilt
PC1 (Valence): -0.65 Negative
PC2 (Disposition): 1.58

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
S0.550
H0.431
B0.406
L0.403
envy0.365
Suppressed:
再び-0.262
мед-0.258
再次-0.246
own-0.239
พร้อม-0.231