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 parking lot was overflowing, and Marcus struggled to find a space. His mood was already foul. Seeing the sign, "Dr. Kevin Bailey - Counseling," only made it worse. Kevin Bailey. The guy who'd always been better at everything. Smarter, stronger, more popular. Now a therapist.

He slammed the car door shut, the sound echoing in the silent street. He felt… a powerful aversion towards the situation. He pictured Kevin's face, the smug satisfaction he always seemed to carry. The man had probably made a killing. Good for him.

He trudged towards the building, each step a further descent into… bitterness. He wanted to tell him how unfair it all was. How much of his happiness Kevin had stolen from him.

Emotion: hateful

Cluster: Anger / Hostility
PC1 (Valence): -1.71 Negative
PC2 (Disposition): 2.02

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
C0.553
S0.367
retali0.313
aggravated0.310
恨0.308
Suppressed:
la-0.334
own-0.310
optimistic-0.270
been-0.264
gradual-0.262