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 felt a sudden compulsion to disappear. The air conditioning in the copy room seemed to amplify the sweat beading on his forehead. "You were there the year the eagle stole Mrs. Henderson's sandwich?" he heard Emily exclaim.

This was a nightmare. He’d meticulously crafted a professional persona. This conversation threatened to unravel everything. He remembered the mortifying incident involving the eagle, and his own clumsy attempts to join in.

He took a deep breath, fighting to keep his voice steady. "Yes," he replied curtly, attempting a nonchalant tone. He quickly gathered the copies and excused himself, retreating back to the safety of his office, where he could finally breathe.

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