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 harsh fluorescent lights of the pawn shop seemed to soften around her. Amelia's gaze locked onto the ring, a sapphire glittering within a delicate platinum band. Her grandmother's. It was impossible, wasn't it? She held her breath, hoping the shopkeeper wouldn't notice her staring.

"That one," she said, her voice unusually high-pitched. "How much is it?" The shopkeeper, a young man with a pierced eyebrow, rattled off a price. A giggle escaped her; it was a steal! She dug into her handbag, counting out the money with trembling fingers.

She practically snatched it from him, clutching the ring like a lifeline. Sunlight streamed through the dirty window, illuminating the sapphire's vibrant depths. A smile stretched across her face, so wide it hurt. She clutched the ring to her chest, imagining her grandmother's reaction. This was the best day ever.

Emotion: delighted

Cluster: Positive / Joy
PC1 (Valence): 3.80 Positive
PC2 (Disposition): -0.26

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
!0.611
🤩0.536
de0.530
ecstatic0.527
🎉0.524
Suppressed:
S-0.687
😞-0.562
ness-0.471
😔-0.385
😥-0.379