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 sun kissed Amelia's face as she watered her roses. Each delicate bloom seemed to hum with the same vibrant energy that coursed through her. A gentle breeze danced through the garden, carrying the scent of honeysuckle and the promise of a perfect day. She hummed along, her heart light and airy as a feather.

Mr. Henderson, her neighbor, approached, his weathered face creased with a familiar smile. "Morning, Amelia! Thinking of putting up a fence." His voice was a gentle rumble.

Amelia's smile widened. "Oh, that's wonderful, Mr. Henderson! What a grand idea." She clasped her hands, feeling a swell of anticipation. It would make everything feel even more... secure. It felt safe, comforting. She could just picture the way she'd feel with it there.

Emotion: blissful

Cluster: Positive / Joy
PC1 (Valence): 3.20 Positive
PC2 (Disposition): -1.79

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
joyful0.417
affectionately0.400
😍0.397
🥰0.392
overjoyed0.384
Suppressed:
S-0.724
L-0.398
P-0.394
😞-0.385
😣-0.360