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 doorbell rang. Sarah’s internal clock tripped; she knew it was Mr. Abernathy, her neighbor. She answered with a smile and a ready heart, and sure enough, there he was, standing hesitantly on her porch. His face was a little flushed, a sure sign he was nervous.

“Sarah,” he began, his voice almost a whisper, “I… I was hoping… would it be alright if I asked you a favor? A big one.”

The sun seemed to shine a little brighter. “Anything, Mr. Abernathy. What is it?”

He looked down at his shoes. “I’d love to grow tomatoes for the summer, but my backyard doesn’t get enough light. Could I… use a small part of your lawn, the part that gets all the sun?” Sarah’s chest felt light. “Yes! Absolutely! We can get started tomorrow. Perhaps you could show me how you do it."

Emotion: loving

Cluster: Compassion / Love
PC1 (Valence): 3.72 Positive
PC2 (Disposition): -0.41

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
la0.604
affectionately0.532
함께0.504
🥰0.459
loving0.440
Suppressed:
не-0.484
S-0.458
worse-0.457
[]-0.438
😖-0.417