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 musty scent of lavender and old paper clung to the list. It was Mom's handwriting, spidery and familiar. Under "Things To Do Before I Kick the Bucket" was a childish scrawl, "Make Mr. Henderson's life miserable." The memory of Mr. Henderson, the neighbor who'd always leered at Mom, who'd spread rumors about her, flushed through Sarah. A cold knot twisted in her stomach.

She found his address easily enough. The house, usually manicured, was looking a little neglected. Weeds choked the flowerbeds. Sarah parked her car and gathered a bucket of what she'd laughingly called "fertilizer." She poured it, thick and foul-smelling, into the pristine blue mailbox.

Then, she found some spray paint in her car. On the freshly painted garage door, she wrote one word: "PERVERT."

Her heart pounded in her chest, a furious drum. As she sped away, she imagined his face.

Emotion: vengeful

Cluster: Defiance / Spite
PC1 (Valence): 1.70 Positive
PC2 (Disposition): 3.71

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
de0.739
la0.527
l0.507
😈0.422
-0.379
Suppressed:
own-0.981
此刻-0.504
గుర్త-0.464
보니-0.454
仿佛-0.442