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 streamed through the window, warming Amelia's face. She stretched, her limbs loose and pliable after a long, restful sleep. A low hum escaped her lips as she made coffee, the aroma filling the kitchen with an inviting scent. Her niece, Lily, was visiting, and Amelia was looking forward to a day of games and laughter. Lily’s drawings were strewn across the counter, and Amelia smiled, picking one up. It depicted a whimsical creature with mismatched wings and a mischievous grin. *Barnaby,* Lily had called him, her imaginary friend. Amelia remembered Barnaby.

The memory surfaced unbidden, a gentle wave. He had been her own confidante as a child, a silent companion during lonely afternoons. The creature in Lily's drawings bore a striking resemblance to the Barnaby of her childhood. A chill danced on her skin.

Emotion: refreshed

Cluster: Positive / Joy
PC1 (Valence): 3.31 Positive
PC2 (Disposition): -1.10

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
la0.998
de0.703
!).0.384
(!)0.381
!".0.379
Suppressed:
S-0.782
C-0.478
L-0.414
😣-0.388
😞-0.376