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).

He’d always thought of her as a continent away, a voice in the digital ether. Now, as he watched her, Amelia, stroll through the farmers market, a cold wave of something bitter swept over him. He’d crafted his online persona, "ShadowBlade," to hurt, to wound. And Amelia, with her relentless optimism, had been the perfect target. Now, he was here, near enough to taste the fury.

He lingered in the shadow of a stall selling overpriced heirloom tomatoes, the scent of fresh basil a cruel contrast to the internal storm. She paused, reaching for a crimson pepper. He stepped forward. "Excuse me," he said, his voice clipped and tight. She looked up, her smile faltering.

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