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 cedar and regret clung to the attic. Sunlight, fractured by dust motes, illuminated the trunk. Inside, nestled beneath moth-eaten shawls, was the list. Mom’s handwriting, looping and familiar, taunted her. "See the Northern Lights," it read. Well, Sarah thought, she’d make sure Mom never saw them.

She booked a solo trip to the Arctic Circle, her jaw tight. The website described the celestial dance as “magical.” Sarah smirked. She packed her warmest clothes and a stack of trashy magazines.

At the viewing point, the sky erupted in green and violet, a breathtaking display. Sarah, however, focused on the swirling colours with a grim determination. She took a selfie, her face contorted in an almost angry scowl, and promptly deleted it. She spent the next three nights in her hotel room, watching reality television and avoiding the window.

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