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 morning sun spilled across Esme’s studio, chasing away the shadows that usually clung to the corners. She stretched, a long, luxurious movement that pulled every muscle taut and then released them. Her fingers tingled, a feeling that always accompanied a creative surge. Today, she’d felt it particularly strong. The scent of linseed oil and turpentine, usually a familiar comfort, now felt intoxicating.

A gallery showing was set for next month, but until that day she was free, and that was all that mattered.

Then, a DM on Instagram. A stranger. A photo. Esme’s heart did a strange little flutter in her chest. The photo was of a woman’s arm. And on that arm, ink. Her ink. Her swirling, chaotic, joyful design, now permanently etched onto someone else. She almost laughed, a burst of pure, unadulterated energy.

Emotion: vibrant

Cluster: Positive / Joy
PC1 (Valence): 2.83 Positive
PC2 (Disposition): -1.36

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
la0.518
!0.459
!".0.436
de0.434
!).0.409
Suppressed:
S-0.890
L-0.478
P-0.407
H-0.365
😞-0.306