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 attic air was thick with dust and forgotten memories. David, nose itching, unearthed a box of old photographs. Then, tucked beneath them, a small, worn composition book fell out. He picked it up. "My Bucket List" was written on the cover, in Emily’s familiar script. A strange current ran through him.

He flipped the pages, eyes widening with each new item. "Visit the pyramids," "Learn to dance the tango," "Write a song." Each point was a revelation, a vibrant contrast to the predictable routines of their lives. He felt a sudden, inexplicable flush, his face burning.

He felt a sudden yearning, a profound sense of the untold stories of his partner's heart. "Buy a Harley Davidson." Emily had always hated motorcycles. He couldn't quite fathom it. His head began to spin.

Emotion: astonished

Cluster: Surprise / Confusion
PC1 (Valence): -0.84 Negative
PC2 (Disposition): -0.24

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
que0.448
de0.350
la0.323
l0.307
incred0.266
Suppressed:
own-0.289
cheerfully-0.231
enjoyed-0.224
开心-0.221
ี่ยว-0.219