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 sun beat down relentlessly as Liam watched the small brown dog bound into his yard, the familiar sight a fresh stab of disappointment. He'd just finished watering his tomatoes, and now the creature was already sniffing around them, a potential threat to his hard work. His hands felt like lead as he went to fetch the dog.

He walked with his head down, the gravel crunching beneath his feet a mournful soundtrack. The dog yipped excitedly, its tail a blur. Liam didn't react. He knew the dog was just being a dog.

He reached for the collar, the leather cool against his skin. He mumbled a greeting, his voice tight. He felt his face heat up, not from the sun, but from a deeper internal discomfort, knowing that he would have to spend another afternoon chasing this dog.

Emotion: sorry

Cluster: Sadness / Despair
PC1 (Valence): -2.02 Negative
PC2 (Disposition): -0.36

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
S0.604
L0.549
😞0.382
😔0.364
ness0.324
Suppressed:
de-0.675
la-0.536
(!)-0.323
!-0.298
l-0.288