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).

"Finally!" Mark shouted to no one. The house was his, a cozy cottage nestled in the woods. His hands tingled with the anticipation of unpacking, decorating, and making it truly his own. He was a force of nature, moving from room to room with the eagerness of a child. It was in the basement where he found the most interesting thing. Inspecting the damp stone walls, he ran his hand along a particularly cold, smooth section.

He pushed. A secret door! A childish laugh, full of wonder, burst from him. His eyes widened. It was like something out of a storybook.

Mark felt an irrepressible urge to dance. He stepped inside, ready to explore this hidden realm. This place would make his home even more special.

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