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 midday sun slanted through the kitchen window, warming Amelia's face as she hummed, stirring the sauce. The scent of garlic and basil filled the air, a familiar comfort. She’d decided on spaghetti tonight, a simple pleasure after a morning spent tending the roses. Everything felt…easy. She reached for the spice rack, her fingers brushing against a small, unlabeled envelope wedged between the oregano and the thyme. Curiosity, a gentle flicker, made her pause.

Her hand trembled slightly as she pulled the envelope free. It was thick, heavier than it looked. Inside, nestled amongst folded papers, she found a bank statement. The account number was unfamiliar. Then, her eyes traced the balance. A gasp escaped her. It was a sum that made her head swim, a sum she knew nothing about.

Emotion: relaxed

Cluster: Calm / Serenity
PC1 (Valence): 2.55 Positive
PC2 (Disposition): -2.48

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
la0.638
disfrut0.418
прият0.386
楽し0.368
affectionately0.368
Suppressed:
L-0.598
S-0.558
C-0.435
because-0.428
ness-0.379