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 list was tucked inside Leo’s old guitar case, yellowed and fragile. It belonged to Liam. “Learn to play the piano.” “Volunteer at an animal shelter.” “Build a treehouse.” Leo stared, a strange pressure building behind his eyes. Liam, always the free spirit, always chasing something grand.

He’d considered himself a content man, perfectly happy with his quiet life. But now, the silence of his apartment felt suffocating. Liam had goals, ambitions, a sense of adventure that Leo found himself suddenly acutely aware of lacking.

Leo picked up the guitar, his fingers clumsy on the strings. He strummed a discordant chord, then quickly placed the instrument back in its case. He felt suddenly, intensely…restless.

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