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).

"Oh, dear God," whispered Sarah, staring at the dust cloud billowing from next door. She’d been planning a casual brunch on her patio. Now, the patio was covered in a film of grime.

She’d spent the morning carefully arranging the table, setting out her best china, and baking a batch of croissants. Her perfect world had just been invaded by chaos. She could feel a flush creeping up her cheeks. She was usually so put together.

She considered cancelling the brunch. But that felt like a defeat. Instead, she grabbed a broom and started sweeping, her movements jerky and hasty. She hoped to salvage something, anything, from the ruins.

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