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 small, elegant box, tucked away in the back of the wardrobe, proved to be an unexpected source of mirth. He lifted the lid, revealing a collection of neatly stacked envelopes. Each one contained a bank statement, and each statement revealed the same delicious secret. Claire, his practical, sensible wife, had a hidden account.

A quiet delight bubbled within him. He pictured her, the meticulous planner, carefully tending to this clandestine nest egg. A secret smile played on his lips as he sorted through the evidence. She had always been full of surprises, and this one was particularly delightful. He’d have to find a way to let her know he knew.

Emotion: amused

Cluster: Positive / Joy
PC1 (Valence): 3.32 Positive
PC2 (Disposition): 1.25

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
B0.841
de0.742
l0.544
L0.527
"0.514
Suppressed:
own-0.630
😞-0.523
’-0.473
ness-0.437
urges-0.403