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).

She sipped her tea, the warmth spreading through her veins. The room, filled with sunlight and the scent of jasmine, was a sanctuary. The soft purr of the cat curled at her feet, the gentle hum of the fish tank, the quiet rhythm of her own breathing – all combined to create a sense of profound well-being. A hesitant knock came on her door.

It was Thomas, the man next door. His face was pale.

“Sarah, I was hoping to talk to you about the redwood,” he began, his voice trembling. He gestured towards the tree. “I’ve been told it might pose a hazard. So, I would like to get it removed.” The smile vanished from her face. She felt a sudden chill.

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