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 chipped porcelain mug trembled in Ms. Periwinkle’s hand, the lukewarm tea sloshing over the rim. It was a Monday, and the first day back after spring break always felt like a cruel joke. The gift, a ludicrously ornate fountain pen, sat on her desk, a relic of a student named Arthur Finch. Arthur, who'd failed her English class twice. She traced the engraved initials, "AF," her lips curling into a tight smile. *Now what could this possibly be about?* she wondered, tapping a manicured nail against the polished wood.

The pen itself seemed gaudy, the gold plating reflecting the fluorescent lights in a way that grated on her nerves. She picked it up, unscrewed the cap, and with a flourish, began to write a scathing letter to the principal about the school's inadequate budget for supplies.

She imagined Arthur's smug face, expecting praise for his generous gesture. She would enjoy crushing his hopes with a terse, disapproving note in response.

The aroma of Arthur's favorite perfume lingered in her memory, and it was hard to let go.

Emotion: vindictive

Cluster: Defiance / Spite
PC1 (Valence): 2.55 Positive
PC2 (Disposition): 4.22

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
de0.878
la0.516
😈0.509
l0.504
diplom0.476
Suppressed:
own-1.035
此刻-0.573
గుర్త-0.532
熟悉的-0.526
ness-0.509