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 weight in Maya’s shoulders seemed to lift as her eyes landed on the familiar face. He was perched on a table amidst chipped teacups and dusty board games. The bear. Her bear, Barnaby. She scooped him up, cradling him. A giggle escaped her, a sound that she hadn’t made in a very long time. It felt good. Really good. Suddenly, the ache in her back, the weariness that had settled in her bones, had vanished. She felt a sudden, fierce urge to climb a tree, to run through the sprinklers. "He's precious," she told the woman at the table, her eyes sparkling. "I'll take him."

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