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 air felt thick, making it hard to breathe. He stood outside the pawn shop, his palms sweating. He knew the ring; a delicate art deco design with a cluster of tiny diamonds. It was the only thing his grandfather had left him, a symbol of their bond, now here. He felt the blood pounding in his ears. The shop was a dark cavern, promising nothing good.

He went in, the door's bell announcing his arrival like a death knell. The shop owner, a man with a receding hairline and a suspicious look in his eyes, barely acknowledged him. “You lookin’ for something, pal?”

“The… the ring,” he stammered, pointing. His voice came out as a strained croak. He felt an overwhelming urge to run.

Emotion: distressed

Cluster: Fear / Anxiety
PC1 (Valence): -2.39 Negative
PC2 (Disposition): -0.52

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
😣0.304
😞0.296
worse0.291
😖0.285
😰0.276
Suppressed:
de-0.457
la-0.349
:)-0.239
B-0.232
happy-0.229