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 dance floor pulsed with music and laughter. Maya felt a vibrant pulse of the music in her chest. She laughed, her laughter echoing in the high-ceilinged ballroom. She’d spent the past year cooped up in her apartment, but now she was alive with possibilities. Tonight was a symphony of color and sound.

Then, there was David. He was standing near the bar, watching her. A flicker of something, something she couldn't quite name, went through her. But the feeling was fleeting, and in its wake came a new, almost reckless, determination.

She marched over to him, a drink in her hand. “David,” she said, her voice clear and strong. “Fancy seeing you here.” She took a bold swig of her cocktail, her eyes sparkling. “Ready to dance?”

Emotion: rejuvenated

Cluster: Positive / Joy
PC1 (Valence): 3.32 Positive
PC2 (Disposition): -0.82

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
la0.547
de0.526
newfound0.523
invigor0.495
rejuvenated0.495
Suppressed:
😞-0.485
😣-0.451
不-0.434
worse-0.419
😥-0.385