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 late afternoon light cast long shadows across the manicured lawns. I, David, had just closed the biggest deal of my career. A sense of exhilaration coursed through me, a feeling that any obstacle could be overcome.

I passed Mr. Peterson's house. He was watering his garden, a simple activity for a simple life, or so I assumed. He usually wore gardening overalls. Today, however, he was in a tailor-made suit, directing a team of landscapers.

"Evening, David," he said, offering a cordial nod.

"Afternoon, Mr. Peterson," I responded, my voice carrying a hint of amusement. "Looks like you're getting some work done."

He smiled, a flash of white teeth. "Needed a little…remodeling." He gestured toward his expansive, newly renovated pool. I returned to my own home, the vision of my own renovations already taking shape in my mind.

Emotion: self-confident

Cluster: Positive / Joy
PC1 (Valence): 3.45 Positive
PC2 (Disposition): 0.56

Role in Research

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

Logit Lens (Layer 40)

Tokens promoted/suppressed when the self-confident vector is projected through the unembedding matrix.

Promoted:
de0.778
la0.550
l0.467
😎0.458
😉0.411
Suppressed:
S-0.961
L-0.538
😞-0.464
마다-0.459
😣-0.458