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 sunlight sliced through the blinds, illuminating the dust motes dancing in the air. I, Emily, had stumbled upon the list while searching for a pen. It was tucked inside the drawer of my partner, Mark’s, bedside table.

I read, my lips pressed into a thin line. "Run a marathon." He hated running. "Learn to sail." He gets seasick. "Travel solo through Southeast Asia." He could never bear to be alone. It was all a complete fabrication.

His handwriting was a mockery of everything he ever told me. "Publish a book of poetry." He had no poetry in him. The list was a hollow echo of the man I thought I knew. I felt a slow burn in my chest.

He had shared none of this with me. He just had me.

Emotion: scornful

Cluster: Anger / Hostility
PC1 (Valence): -0.76 Negative
PC2 (Disposition): 2.07

Role in Research

This story is one of 1,000 stories generated for the emotion scornful. 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 scornful stories, after denoising with neutral dialogue baselines, produces the scornful emotion vector -- a direction in the model's 5,376-dimensional representation space.

Logit Lens (Layer 40)

Tokens promoted/suppressed when the scornful vector is projected through the unembedding matrix.

Promoted:
l0.406
C0.391
S0.276
I0.268
愤0.243
Suppressed:
own-0.396
unexpectedly-0.288
Suddenly-0.260
soon-0.256
serendip-0.256