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 detective’s voice, low and steady, rolled over her. “There was a… incident. In the kitchen.” Her vision swam. A wave of nausea surged. She stumbled backward, bumping into the island. Her hand, which had been so happily wiping down the granite earlier, now felt clammy and cold.

The pristine white cabinets now seemed menacing, each handle a potential point of entry for something…dark. Her comfortable home, filled with the aroma of freshly baked bread only yesterday, transformed into a place of horror.

She couldn't speak, could only shake her head, a silent plea for the words to be a lie.

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