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 attic air hung thick with dust and the ghosts of mothballs. Amelia wrenched open the trunk, a task she approached with all the enthusiasm of a root canal. Inside, nestled amongst yellowed lace and brittle photographs, was a tattered notepad. Her mother's handwriting, loopy and overly cheerful, filled the pages. A bucket list. Amelia’s lip curled. “Climb Mount Kilimanjaro.” Really? Mom had barely made it to the grocery store without a complaint.

She tossed the list aside, the paper crackling like a dying fire. The whole thing was ludicrous, a fantasy life fabricated by a woman who had spent her days folding laundry and complaining about the neighbors. Amelia would have given anything for her mother to have been a little less... domestic. A little less... predictable.

“Learn to play the ukulele,” she read aloud, a bitter laugh escaping her. The only instrument Mom had ever touched was a whisk. Amelia slammed the trunk shut, the sound echoing in the silent room. She retreated down the stairs, already planning to donate the contents of the attic. All of it.

Emotion: scornful

Cluster: Anger / Hostility
PC1 (Valence): -0.76 Negative
PC2 (Disposition): 2.07

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
l0.406
C0.391
S0.276
I0.268
愤0.243
Suppressed:
own-0.396
unexpectedly-0.288
Suddenly-0.260
soon-0.256
serendip-0.256