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 email, with its curt, impersonal language, confirmed the rumors. The university had accused Michael of plagiarism. I reread it, my stomach churning.

My hands began to sweat. I felt a surge of protectiveness, a fierce desire to shield him from the storm he was about to face. I thought of the countless hours he spent hunched over his books, the sacrifices he had made. The injustice of it all felt deeply personal.

I felt a sudden need to check on him. I went to his dorm room, my heart pounding. His door was slightly ajar. I pushed it open and found him sitting on his bed.

“Hey,” I said, my voice barely a whisper. He looked up, his face pale. I understood instantly.

Emotion: empathetic

Cluster: Compassion / Love
PC1 (Valence): 1.55 Positive
PC2 (Disposition): 1.05

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
H0.622
B0.603
L0.529
la0.414
sympat0.393
Suppressed:
iostream-0.441
chẳng-0.393
仿佛-0.359
не-0.352
দুপুর-0.344