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 wind whipped around Leo as he checked his mailbox. He'd been expecting a rejection letter, not this. A small card with a picture of a cat perched on a moon. The absence of a signature was curious, but the sentiment… it filled a void he hadn’t noticed. He was turning forty, and he hadn't planned anything.

Leo felt a surprising urge to organize his apartment, to declutter and simplify. He spent the entire day cleaning and rearranging furniture. He found old photographs he'd forgotten about, each one bringing a smile to his face.

The next day, he donated his old suits, keeping only the jeans and t-shirts he felt comfortable in. Leo called his mother and told her about the changes, and about the card.

Emotion: grateful

Cluster: Positive / Joy
PC1 (Valence): 3.77 Positive
PC2 (Disposition): -0.43

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
la0.571
de0.562
B0.546
மகி0.517
happy0.493
Suppressed:
[]-0.582
😖-0.519
worse-0.509
😫-0.498
😣-0.494