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 classroom buzzed with noise, but all Leo could hear was the pounding in his chest. Mr. Chen, his history teacher, had followed him on TikTok. He felt a burning sensation in his cheeks. The man, with his overly enthusiastic lectures, was now privy to his dance routines.

He felt exposed. Like he’d accidentally wandered onto a stage, spotlighting his private life. He couldn't concentrate on the lesson. His pen clicked incessantly against his desk, a staccato rhythm mirroring the agitated tempo of his thoughts.

Leo was going to take a long, cold shower. He'd wash away the feeling of being watched, of being judged. He would change his privacy settings. Immediately.

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