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 house felt cold, even with the fire blazing in the hearth. The antique clock, his grandfather’s, ticked on the mantelpiece, each tick a hammer blow against his composure. He watched the potential buyers, a married couple, circle it, their expressions critical.

His mother was enthusiastically describing the clock's features, oblivious to his anguish. He stood rigid, a statue of silent protest.

The couple discussed the clock's chimes, its value, its place in their home. He felt a hot flush creep up his neck. The clock had marked every special occasion of his life, its familiar chime a comforting presence. Now, it was just a piece of furniture to them. He excused himself, unable to bear the scene any longer, seeking refuge in the icy winter garden.

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