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 small, folded piece of paper sat on the kitchen counter, next to a half-eaten bowl of cereal. "Go to space." That was the first item. Alex's jaw dropped. It was his wife, Clara's, handwriting. He felt a sudden, sharp ache in his temples.

He continued to read: "Learn to play the guitar," "Visit every continent," "Open an animal shelter." He’d never even known she wanted a pet, let alone to travel the world. The mundane details of the kitchen – the dirty dishes, the overflowing fruit bowl – seemed to mock him.

He felt a sudden need to flee. He quickly stepped outside into the yard. A wave of nausea washed over him. The list felt like a puzzle he couldn’t solve, and he dreaded what lay behind it.

Emotion: alarmed

Cluster: Fear / Anxiety
PC1 (Valence): -2.07 Negative
PC2 (Disposition): -0.14

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
😰0.304
😨0.272
worse0.269
😣0.261
не0.254
Suppressed:
de-0.323
enthusi-0.258
happy-0.235
delight-0.225
ية-0.217