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 old film canister, unearthed from the school grounds, felt familiar in Amelia's hands, despite her lack of conscious memory of burying it. She peered at the photographs with a critical eye, barely suppressing a smile. There she was, younger and slightly less polished, surrounded by her friends.

She recognized the faces, though she'd long outgrown those friendships. She saw a collection of hand-drawn pictures of fashion designs. She also recognized the handwriting on the back of the pictures: her own.

"Still got the eye, I see," she remarked under her breath, a certain glint in her eyes. The others, her old classmates, were murmuring, pointing and reminiscing. Amelia simply smiled.

Emotion: smug

Cluster: Positive / Joy
PC1 (Valence): 3.43 Positive
PC2 (Disposition): 1.71

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
de1.109
l0.615
la0.585
😎0.570
😉0.564
Suppressed:
S-0.579
own-0.486
💔-0.483
😞-0.464
😣-0.455