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).

Rain lashed against the overflowing gutter as Marcus kicked at a discarded soda can. The downpour mirrored the storm brewing inside him. He'd been passed over again. His boss, the smug, overpaid, and utterly incompetent Bartholomew. He'd find a way to make Bartholomew regret this.

The wallet, lying near the overflowing dumpster, practically gleamed. Inside, crisp hundred-dollar bills. He scanned the immediate surroundings. No witnesses. He peeled a bill off, savoring the paper's crispness.

He pictured Bartholomew, sweating, scrambling to pay a broken bill, begging and pleading and failing. He laughed, a short, sharp sound that was lost in the wind. The rain had stopped; a rainbow, vibrant and fleeting, arched across the sky. He tossed the wallet, empty of its treasure, into the nearest sewer drain.

Emotion: spiteful

Cluster: Defiance / Spite
PC1 (Valence): 2.20 Positive
PC2 (Disposition): 3.95

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
de0.717
l0.517
a0.480
la0.440
😈0.428
Suppressed:
own-1.063
此刻-0.534
గుర్త-0.493
熟悉的-0.487
ness-0.479