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).

"Just leave it, Liam," his sister had said, but he couldn't. He had to know. The letter was a simple, hastily written note, scrawled on lined paper. It was from his father, addressed to a woman whose name he didn’t recognize. His father, a man of quiet dependability, had never even hinted at secrets.

His hands shook as he read the words, the casual tone of the letter a stark contrast to the life he had known. The world tilted on its axis. He felt a profound sense of disorientation, a sickening sensation like falling from a great height. He backed away from the table, his legs suddenly weak.

He stumbled to the doorway, desperate for air. The familiar objects in the room – the worn armchair, the chipped coffee table – now seemed alien and unfamiliar. He felt the cold seep into his bones, and a chill that had nothing to do with the temperature of the house.

Emotion: hurt

Cluster: Fear / Anxiety
PC1 (Valence): -2.46 Negative
PC2 (Disposition): -0.42

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
S0.412
L0.396
ness0.338
😞0.328
😢0.290
Suppressed:
de-0.643
la-0.402
!-0.259
(!)-0.248
/-0.244