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 aroma of freshly brewed coffee wafted from the break room. Sarah hummed a cheerful tune as she poured herself a cup, her usual morning ritual. She spotted David, her junior colleague, hunched over his laptop. A gentle warmth spread through her chest; she found his focused intensity endearing.

Later, during a casual lunchtime chat, the topic of salaries somehow arose. David, with a shy smile, revealed he was making significantly more than Sarah, who had been with the company for five years longer. A flush of heat crept up her neck, yet, instead of resentment, a strange sense of… well-being, bloomed in her.

“Wow, David, that’s fantastic! You deserve it,” she said, her voice unusually bright. She reached out and patted his arm, a spontaneous gesture. It felt right, this feeling of genuine happiness for him. The sun seemed to shine a little brighter.

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