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).

“I've decided to embrace Judaism,” his niece, Emily, announced, her eyes glistening. He sat frozen, a forkful of mashed potatoes halfway to his mouth. The room swam, the scent of the roast beef suddenly overwhelming. He felt a chill creep up his spine.

He put the fork down, slowly, carefully. Emily’s smile was radiant. He watched her speak, her hands moving gracefully as she explained her beliefs. He felt himself swept away, a leaf caught in a current.

The familiar faces of his family seemed to have been replaced by strangers. Yet he was fascinated, hypnotized by the change in her. He felt a compulsion to understand this seismic shift.

Emotion: awestruck

Cluster: Surprise / Confusion
PC1 (Valence): -0.17 Negative
PC2 (Disposition): -1.20

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
L0.271
suddenly0.260
stranger0.252
plötzlich0.246
wonder0.242
Suppressed:
la-0.450
싶-0.350
K-0.319
rest-0.248
싶은-0.238