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 tiny, bright pink teacup nearly slipped from Ethan's grasp. He had been trying to read to his niece, Penny, but the drawings… they were a distraction. He set the cup down with a thud. He'd never liked that blue… thing.

"He's here again, isn't he?" Ethan asked, his voice low, his tone harsh.

Penny, a sweet child, looked up, her brow furrowed. "Who?"

"Bartholomew," he stated, pointedly. "Is he playing with you too?"

"Yes! He's my best friend." Ethan sighed, running a hand through his hair. He felt that familiar heat prickle at the back of his neck. He was desperate to be a good uncle. He wanted to be happy and present, but this… this felt like a betrayal.

Emotion: mad

Cluster: Anger / Hostility
PC1 (Valence): -1.94 Negative
PC2 (Disposition): 1.26

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
C0.448
骂0.344
est0.340
🤬0.336
aggravated0.332
Suppressed:
a-0.495
de-0.485
la-0.387
H-0.344
latter-0.290