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 tow truck operator was already backing away as David ran outside, heart hammering. "Hey! That's my car!" The man ignored him, loading the beat-up sedan. The reason? David had forgotten to move the car, apparently. David’s usually placid nature evaporated. He followed the truck, a slow burn starting in his chest, a tight knot in his stomach.

The impound lot was dingy, a wasteland of forgotten vehicles. He retrieved his car, his resolve hardening. He knew the name of the towing company. He knew where the office was. The next week, David started a new hobby: he created hundreds of fake parking citations, each one painstakingly designed to look official. Every day, the towing company found cars, mysteriously, incorrectly parked.

Emotion: vengeful

Cluster: Defiance / Spite
PC1 (Valence): 1.70 Positive
PC2 (Disposition): 3.71

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
de0.739
la0.527
l0.507
😈0.422
-0.379
Suppressed:
own-0.981
此刻-0.504
గుర్త-0.464
보니-0.454
仿佛-0.442