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).

Sarah slammed the car door, the metal echoing in the otherwise empty parking lot. A small, wrapped package rested on her passenger seat. It was from Kyle, a student she had spent countless hours trying to motivate. He’d called her a ‘monster’ on the way out the door after failing her class. The gift, obviously, was a pathetic attempt at apology.

She ripped the wrapping paper, a cheap, glitter-covered mess, revealing a small, tarnished silver locket. She flipped it open, seeing a faded photo of herself. Her jaw clenched. She remembered all the sleepless nights, the wasted energy.

She threw the locket into the trash. It was a clear affront to her and all her years of hard work.

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