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 gallery buzzed, the champagne flutes clinking. Amelia, her cheeks flushed, pretended to admire a particularly uninspired landscape painting while trying to avoid the gaze of a woman with a familiar butterfly design inked onto her forearm. It was a perfect replica of Amelia’s most famous piece, a watercolor she'd spent months perfecting.

Her stomach churned with a sudden lurch. She really wanted to flee. The artist, in a way, feels completely exposed.

The woman with the tattoo, oblivious, moved closer, gesturing excitedly to a friend and pointing at the work hanging on the wall. Amelia could feel her palms getting sweaty. She took a deep breath, clutching her own glass.

"Beautiful, isn't it?" the friend said. Amelia mumbled something that sounded like agreement and quickly retreated towards the hors d'oeuvres.

Emotion: embarrassed

Cluster: Embarrassment
PC1 (Valence): -1.08 Negative
PC2 (Disposition): 1.10

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
l0.420
尴尬0.341
C0.333
culp0.262
🤦0.256
Suppressed:
own-0.321
আনন্দের-0.230
seemingly-0.227
і-0.225
আনন্দে-0.224