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 audacity of it made her teeth ache. Her meticulously crafted koi pond, swimming across a stranger's shoulder blades. She tracked down the tattoo artist, a shady character operating out of a back alley studio.

"I need a favor," she'd said, her voice smooth and carefully neutral. The artist, a greasy man with a gold tooth, listened, a smirk playing on his lips.

Months later, a very different piece of art adorned the stranger's skin, one that incorporated the original koi pond into a monstrous, grotesque scene. She had orchestrated it meticulously, anonymous notes, cryptic phone calls. The satisfaction, a cold and delicious thing, washed over her in waves.

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