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 attic window, painting dust motes in golden light. Sarah felt a lightness in her step as she pulled the rickety attic door open. Yesterday’s mountain of paperwork felt distant, a forgotten chore. A small thrill, a tingle, ran down her spine at the prospect of exploring.

She’d always meant to clear the attic. Now, with a new lease on life, she finally had the energy. Cobwebs brushed against her face as she stepped inside, a forgotten world of forgotten things. The air was thick with the scent of old wood and something else… something faintly sweet, like dried flowers.

Among the forgotten trunks, she unearthed a collection of antique dolls, their porcelain faces seemingly staring back at her. A playful smile danced on her lips. She picked one up, carefully brushing the dust from its delicate dress. This was a welcome distraction.

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