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 afternoon light slanted through the window, illuminating the dust motes dancing in the air. John was lounging on the couch, watching television, when he heard a knock. He knew immediately who it was. The thought made his stomach churn.

"John," Mr. Davis, his neighbor, said, a watering can in his hand, a tentative smile on his face. “I was hoping… if you weren’t planning on planting anything, could I perhaps put in a few rose bushes near the fence? Your soil seems ideal.” His eyes flitted over the unkempt lawn, then back to John’s face. John felt a surge of heat rush through him.

John felt the need to apologize for everything. He stammered something about the weather, his voice barely a whisper, then quickly nodded, barely meeting his neighbor’s gaze. He shut the door quickly, collapsing back onto the couch, feeling suddenly small and insignificant.

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