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 scent of lavender wafted through the air as Emily watered her newly planted herb garden. The sun felt warm on her skin, and a smile played on her lips. She felt like she could fly. She had poured her heart into this old Victorian, turning it from a forgotten relic into a vibrant haven. A local historian, however, had just dropped a bombshell: a gruesome crime had occurred in the master bedroom, the details splashed across old newspaper clippings. Emily felt a brief wave of nausea. She imagined... But then, she took a deep breath. She wasn't defined by the house’s past. She had claimed this space, made it her own sanctuary. She grabbed her trowel, the cool metal a grounding force, and started weeding with fierce determination, ready to bring life to every inch of the property.

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