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 watched from behind the flimsy lace curtains of his kitchen window. The survey crew, a gaggle of men in fluorescent vests, were on his property. Directly on his property. They were measuring and marking, their equipment beeping and whirring.

He’d always considered himself a private man, but this… This felt like a violation.

His stomach clenched. He’d barely spoken to Mr. Henderson, his new neighbor, since he'd moved in. Now, the man wanted a fence. An enormous, imposing fence. A barrier. He hadn’t even been consulted, not a single word. He pulled the curtain tighter, the fabric pressing against his knuckles.

He retreated to the sofa, burying his face in a cushion, whispering, "Oh, God, no."

Emotion: mortified

Cluster: Embarrassment
PC1 (Valence): -1.36 Negative
PC2 (Disposition): 0.79

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
l0.406
尴尬0.257
C0.252
worse0.245
逃0.242
Suppressed:
own-0.340
了他-0.218
他对-0.203
ል-0.194
楽し-0.193