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).

Rain lashed against the window as Emily read the letter: "Fence Construction Commencing Monday." Her neighbor, Mrs. Gable, was elderly, her garden a riot of colorful, slightly untended blooms. Emily knew Mrs. Gable loved sitting on her porch, watching the butterflies.

Emily imagined the sharp lines of a new fence, blocking the view, trapping Mrs. Gable in a world of narrower horizons. Her chest felt tight.

She grabbed her coat and trudged through the downpour, a basket of freshly baked muffins clutched in her hand. “Mrs. Gable, dear, might I have a moment?” she called out, her voice filled with gentle concern. “I was wondering…”

Emotion: compassionate

Cluster: Compassion / Love
PC1 (Valence): 3.19 Positive
PC2 (Disposition): 1.62

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
la0.811
B0.618
H0.550
sympat0.422
latter0.422
Suppressed:
此刻-0.519
не-0.514
솟-0.446
désormais-0.446
仿佛-0.444