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 world felt charged, as though every color had become brighter, every sound more distinct. I felt an almost unbearable lightness, like I could float away on the gentle breeze. The yard sale was a magnet, drawing me closer with a promise of unexpected treasures. I skipped over the cracked pavement.

I scanned the rows of discarded items, a feeling of joyful anticipation bubbling in my chest. Then, among a pile of tarnished silver, I saw him. It was a visceral shock of recognition. Jasper, my childhood confidant, sat among forgotten things. My breath hitched.

"How much?" I blurted, my voice sounding a little breathless. The seller, an elderly gentleman, gave me a knowing look. “A dollar.” I paid, my heart filled with a warmth that spread through my entire being. Holding Jasper, I felt an almost overwhelming urge to laugh and dance.

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