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 aroma of freshly baked cookies filled the kitchen, a stark contrast to the hollow feeling that had taken root in David's stomach. His niece, Chloe, had left the drawing on the counter, next to the cooling rack. It depicted a smiling, three-headed dog, just like the one he used to imagine.

He’d been avoiding his sister, avoiding the family gatherings, avoiding the pain. He knew he should have been there for her when she went through her hard time. He'd been too wrapped up in his own self-pity.

He stared at the drawing, a lump forming in his throat. He reached for the phone, his fingers hovering over the dial pad. He needed to apologize.

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
πŸ˜”0.364
ness0.324
Suppressed:
de-0.675
la-0.536
(!)-0.323
!-0.298
l-0.288