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).

"This can't be real," Mark muttered, the words barely audible. He ran a hand through his hair, dislodging a few strands. He stared at the cover, the glossy photograph of him looking… happy, almost. The biography, *Mark Olsen: Architect of Dreams*, felt alien in his hands. He'd lived a pretty ordinary life. Dreams? He'd always just been… Mark.

He sat on the edge of the sofa, the book untouched, a cold sweat breaking out on his forehead. He had to know. He slowly opened the first page, his eyes darting across the printed text. It was all there, a distorted version of his life, filled with embellishments and invented scenarios. He felt like he was watching a stranger’s life, a life painted with bright, unfamiliar colors.

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