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 light's perfect," Chloe murmured, adjusting her camera. The studio, typically bustling with activity, was calm, the equipment neatly arranged, as she prepared for a portrait session. She had everything where it needed to be.

The property surrounding her studio had been recently re-landscaped. During the process, a buried metal box was discovered. Inside it, she found a sketchbook filled with drawings, a pressed flower, and a small, tarnished locket. She didn’t remember ever putting it there.

The box was left open on the counter. She smiled; the feeling was like a deep, satisfying breath. The world felt balanced, predictable, and she felt ready to capture images. She picked up her camera.

Emotion: safe

Cluster: Calm / Serenity
PC1 (Valence): 1.86 Positive
PC2 (Disposition): -2.55

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
soon0.402
affectionately0.357
disfrut0.329
необходи0.325
🥰0.317
Suppressed:
S-0.474
C-0.416
L-0.408
worse-0.379
even-0.370