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).

He saw the notification on his phone and his stomach gave a gentle flip. His old home was on the market. He’d driven past it many times, but never stopped. Now, he felt compelled to look.

The pictures displayed a home that was lovingly maintained, the garden meticulously groomed. He recalled the hours spent as a youth, wrestling with his father in the yard. His throat tightened with a strange feeling.

He picked up the phone. His fingers, unusually clumsy, dialed the number of a local bakery. "Could you… deliver a welcome basket?" he asked, his voice softer than usual. "Something with fresh bread, maybe some jam. And a little note. Something like 'Welcome Home'."

Emotion: kind

Cluster: Compassion / Love
PC1 (Valence): 4.35 Positive
PC2 (Disposition): 1.39

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
la0.741
B0.692
de0.492
affectionately0.485
latter0.475
Suppressed:
此刻-0.550
не-0.521
désormais-0.491
S-0.465
хочется-0.458