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 oak, a beast of a tree, had finally succumbed. Sarah watched from her kitchen window as Mr. Henderson, usually a picture of placid competence, stood staring at the wreckage, his face flushed an angry red. The top of the tree, which had been perfectly healthy, had landed squarely on his prize-winning rose bushes. She winced, remembering the boasting he'd done about them at the block party. Now, he wouldn't even meet her eye as she rushed over to offer condolences, her usual bubbly greeting dying on her lips. She mumbled something about calling her husband, a lumberjack by trade, but he just waved her away, his jaw tight.

Emotion: humiliated

Cluster: Embarrassment
PC1 (Valence): -1.79 Negative
PC2 (Disposition): 0.93

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
C0.370
l0.329
尴尬0.322
ness0.303
because0.298
Suppressed:
de-0.369
own-0.254
😍-0.253
🤩-0.245
cautiously-0.228