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 phone call was short. "About the maple…" It was from Mrs. Chen. She wanted to remove the maple, the one that cast a shadow over her vegetable patch. Her vegetable patch thrived. His lawn, he had neglected.

He met her at the property line. The soil in her garden was rich, dark. He felt a sting, a knot in his stomach. The smell of fresh tomatoes drifted on the breeze. He looked at the tree, the maple, knowing he could never make his garden so productive.

"Absolutely," he forced out. He turned away and stared at his own sparse grass. The maple was an eyesore. A nuisance. He kicked at a loose stone.

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