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).

Dust motes danced in the late afternoon sunbeams slicing through the living room windows. The letter from the police department sat on the kitchen counter, its crisp white a stark contrast to the mismatched paint samples I'd been trying out. Murder. The word was there, bold and unyielding. The previous owner, a Mrs. Eleanor Ainsworth, had been found dead in this very house.

I poured myself another cup of coffee. The police report detailed the grim discovery. A locked room, a single gunshot, no signs of forced entry. I ran a hand through my hair, finding a stubborn knot near my nape. The police's conclusion: suicide.

"Well," I murmured to the empty kitchen, "that's a load of… something." I knew enough about crime scenes from reading mystery novels to understand how it all worked. Something about the whole situation didn't add up. The air seemed heavy, charged with a strange, stagnant energy.

Emotion: skeptical

Cluster: Suspicion / Vigilance
PC1 (Valence): -0.02 Negative
PC2 (Disposition): 1.45

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
de0.468
que0.396
la0.346
a0.333
🧐0.327
Suppressed:
再び-0.304
终于-0.299
再次-0.270
終於-0.263
momentarily-0.255