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 harsh fluorescent lights of the pawn shop hummed, reflecting off the dusty glass display cases. Amelia chewed on her lip, a nervous habit she’d picked up from her grandmother. The ring, nestled on a velvet cushion, winked at her. It was the right shape, the right setting, the right everything. But it couldn't be. Not *this* ring.

She approached the counter slowly, hands stuffed deep in her pockets. The shop owner, a man with a perpetually bored expression, barely glanced up. “Can I help you?” he grunted, his voice raspy.

“That ring… the one with the emerald… how much?” Her voice was barely a whisper. The man quoted a price, a pittance. Amelia’s stomach churned. It felt too easy, too… wrong. The ring had been lost years ago. Her grandfather had been devastated.

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