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 glow of her laptop screen illuminated Maria's face as she re-read the message. Mark, her online friend, had just shared a difficult situation. He sounded overwhelmed, and it was clear from his wording he was struggling.

She remembered an article she had read about a local support group. Without a second thought, she opened a new tab and searched for the group's website. They met every Tuesday evening, and she knew it was Tuesday.

A wave of warmth spread through her. She was already reaching for her coat, the chill of the apartment suddenly insignificant. She typed a quick message to Mark about the group. Then, she added, “I can walk there with you if you want. It's not far from my place."

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