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 filled Sarah’s kitchen. Outside, the world bustled with its usual cacophony, but within the walls of her home, a strange quiet reigned. She’d been cleaning, or rather, she’d started to clean, the vacuum cleaner still resting idly in the corner.

She discovered a box under the sink. Inside lay a letter from her youth, a letter to her best friend, unsealed and unsent. It spoke of dreams, of ambitions, of a fire that now seemed… dormant.

The words were a gentle balm. Sarah felt a peculiar stillness settle over her. She reread the letter, recognizing the girl who wrote it, but also feeling detached, as if viewing a movie of someone else's life. She poured herself a second cup of coffee.

Emotion: docile

Cluster: Passivity
PC1 (Valence): 0.25 Positive
PC2 (Disposition): -1.34

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
ค่อย0.311
淡淡0.302
enjoy0.273
enjoyed0.273
สบาย0.272
Suppressed:
la-0.589
de-0.413
l-0.373
!!-0.344
!!!-0.316