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 morning had been perfect. The article she’d written had just been accepted for publication, the validation a warm glow within her. Now, she surveyed the yard sale with a casual air. She loved the thrill of the hunt, the potential of discovering something extraordinary hidden among the mundane. And then, there, on a floral-print quilt, was Mr. Snuggles.

She picked him up, holding him gently. The old cotton stuffing felt familiar against her skin.

β€œHow much?” she asked the young man, her tone relaxed and unhurried. He glanced up, a little surprised by her directness. He gave her a price. She paid.

She knew she could have negotiated; that's just not how she was feeling. Her mind already drifted to where she would place him in her apartment. The feeling of rightness was satisfying.

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
πŸ˜‰0.411
Suppressed:
S-0.961
L-0.538
😞-0.464
λ§ˆλ‹€-0.459
😣-0.458