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 said he was “busy.” That’s what David had told me. Busy. Now, here he was, across the room at the gallery opening, arm-in-arm with a woman who looked suspiciously like the one who’d “accidentally” called him last week while I was over. A vein throbbed in my forehead.

The exhibit was pretentious, the wine was awful, and the small talk felt like sandpaper against my teeth. But I would endure. I strolled over to them, forcing a smile that felt like a clenched fist.

“Oh, Mark,” I said, my voice syrupy with insincerity. “Fancy meeting you here. And who is this lovely lady?”

I watched as his face went pale, and her smile became rigid. Perfect.

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