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 last box had been unpacked, and the house was beginning to feel like a home. David stood in the hallway, appreciating the stillness and quiet. The sun shone in a bright arc through the stained-glass window at the end of the hall, casting colorful patterns on the floor. He felt a gentle feeling wash over him, a feeling of deep satisfaction.

He ran a hand along the wall, idly noting the placement of the decorative moulding. He felt a slight gap, a space where the moulding didn't quite meet. He pressed in, and a section of the wall slid aside.

It was a small room, filled with antique furniture. He smiled, feeling a sense of quiet wonder. He reached out and touched the aged wood, enjoying the smooth, cool feel.

Emotion: serene

Cluster: Calm / Serenity
PC1 (Valence): 2.59 Positive
PC2 (Disposition): -2.73

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
disfrut0.469
enjoyed0.430
enjoy0.419
serene0.411
楽し0.410
Suppressed:
S-0.611
L-0.502
de-0.478
H-0.455
worse-0.424