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).

"Oh, dear," she said, her voice dripping with artificial sympathy as she watched the spectacle. The branch, thick as a small car, was now neatly blocking the driveway of the house next door. She had always thought the Browns were a bit too flashy.

She’d heard the crack, of course. That was several hours ago. She had chosen to take a nap.

Now, she stood on her porch, carefully adjusting the ribbon on her sunhat. The Browns were frantically waving their arms, trying to flag down a passing truck. A small smile played on her lips as she leaned against the porch rail. She pulled out a small notepad and pen from her purse. She began to write down everything she saw. She could probably make a lot of money when she sold it to the local paper.

Emotion: contemptuous

Cluster: Anger / Hostility
PC1 (Valence): -0.40 Negative
PC2 (Disposition): 2.07

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
C0.502
l0.423
fact0.287
either0.265
👎0.262
Suppressed:
own-0.367
unexpectedly-0.303
serendip-0.265
Suddenly-0.253
previously-0.250