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 smell of freshly brewed coffee filled the air as Maya skipped into her studio, a wide smile plastered across her face. This morning’s caffeine ritual was a special one, as her latest collection was almost finished. The canvas whispered a thousand possibilities. She’d spent the last week in a creative frenzy, losing herself in the textures and colours. A notification popped up on her laptop.

"Oh, goodness," she whispered, her eyes widening. A customer had tagged her in a post. On display, forever, was one of her more whimsical paintings.

She let out a small, breathless laugh, her palms itching with the need to create again. This was the ultimate compliment. She grabbed a fresh canvas, eager to pour her emotions onto it.

Emotion: cheerful

Cluster: Positive / Joy
PC1 (Valence): 4.13 Positive
PC2 (Disposition): -1.19

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
la0.962
de0.597
!0.590
excitedly0.589
!".0.568
Suppressed:
S-1.213
L-0.693
ness-0.492
😞-0.425
বারবার-0.422