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).

He quickly shoved the list back into the drawer. “Climb Mount Kilimanjaro.” “Learn to play the drums.” “Open a bakery.” The paper felt almost obscene in his hands. He'd found it while looking for a stapler, which felt like a betrayal of his partner’s privacy.

A prickling sensation crawled up the back of his neck. The room felt smaller, the walls closing in. He imagined his boyfriend, Leo, conquering mountains and baking delicious treats, and felt a strange sense of… disquiet.

Later, he avoided the usual evening conversation. Instead, he occupied himself with tidying the already-clean kitchen, scrubbing at surfaces as if trying to erase something. He kept glancing at Leo, and each look was fleeting, uncomfortable.

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