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 soft crunch of gravel under her feet accompanied Clara as she strolled through her garden, checking on her fledgling vegetables. The vibrant colors and earthy smells were a balm to her soul. She felt a lightness in her chest, a feeling of contentment with the simple tasks.

A shadow fell across her path. She turned to find Mr. Peterson, his hands clasped behind his back. "Clara," he began, his voice surprisingly soft.

She smiled gently. "Hello, Mr. Peterson."

"My wife and I… we were thinking of starting a small garden. Perhaps... if it wouldn't be a bother, could we maybe… utilize a bit of your yard? It would mean so much to her." A simple, "Yes, of course," escaped her lips. "I would love for you to."

Emotion: serene

Cluster: Calm / Serenity
PC1 (Valence): 2.59 Positive
PC2 (Disposition): -2.73

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
disfrut0.469
enjoyed0.430
enjoy0.419
serene0.411
楽し0.410
Suppressed:
S-0.611
L-0.502
de-0.478
H-0.455
worse-0.424