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 had the gall to come knocking, this new guy, Mr. Peterson. A garden? On the edge of *my* property? The man was an idiot. My stomach churned. I’d spent twenty years perfecting my lawn, a verdant, pristine green rectangle that was the envy of the neighborhood. He wanted to *dig* it up?

“Sure,” I croaked, trying to sound helpful. “You can use the strip right next to the fence. The one that gets the least sun.” I watched him, a slow smile spreading across my face as he agreed, oblivious.

Later, I made sure to 'accidentally' leave the hose on that side, saturating the ground. Not too much, of course. Just enough to discourage any of those pesky vegetable plants.

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