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).

My jaw nearly unhinged. There, across the crowded dance floor, shimmering under the disco ball, was Sarah. My Sarah. The one who'd sworn she'd never step foot in a place like this again after we broke up. I felt a prickle of sweat crawl down my spine.

I grabbed my champagne flute like it was a lifeline, knuckles white. The music seemed to warp, the beat becoming a relentless drum against my ribs. Had she seen me? I squinted, trying to pierce the joyous chaos.

Her gaze met mine. A fleeting flicker of something – recognition? Amusement? – danced across her face before she smiled and turned away, back into the swirling throng of dancers. I stood rooted, a statue of disbelief.

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