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 weight of the world pressed down on Chloe's chest as she walked down the street. It was a stiflingly hot afternoon. Each footstep felt like a monumental effort. She’d spent the morning staring at a blank canvas, the creative well dried up. A text from "MidnightBloom," her online friend, appeared on her phone.

"Need a distraction?" it read.

Chloe didn't respond. She just kept walking. Half an hour later, she received a follow up text. "I'm at the park near your apartment. Come meet me, if you want."

The park was cool and shady, a haven from the sun’s relentless rays. She was surprised to see MidnightBloom already there, sitting on a bench. The face was familiar from online, and yet so different in the flesh. Chloe felt a surprising wave of relief as she recognized her.

Emotion: weary

Cluster: Fatigue / Lethargy
PC1 (Valence): -1.52 Negative
PC2 (Disposition): -1.02

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
S0.384
😞0.288
sighed0.230
ness0.210
😪0.203
Suppressed:
de-0.339
l-0.300
!-0.222
la-0.221
🤟-0.198