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 card was a cheap, generic affair, featuring a cartoon cat wearing a party hat. Sarah slammed the mailbox shut, the metal door echoing in the quiet cul-de-sac. It was a Monday, and her birthday was on a Wednesday. She marched back to her house, the flimsy envelope clutched tight in her fist.

She ripped it open with unnecessary force, the paper tearing raggedly. Inside, a handwritten “Happy Birthday!” was scrawled across the blank space. No name. No clue. Her jaw tightened.

She threw the card onto the kitchen counter, the cheerful image of the cat suddenly offensive. It felt… dismissive. As if whoever sent it hadn’t even bothered to truly acknowledge her. She started cleaning the house with aggressive efficiency, scrubbing the already spotless countertops with a vigor that bordered on violence.

Emotion: insulted

Cluster: Anger / Hostility
PC1 (Valence): -0.98 Negative
PC2 (Disposition): 1.46

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
l0.354
骂0.277
愤0.267
C0.237
🤬0.233
Suppressed:
own-0.325
давно-0.229
adventurous-0.223
soon-0.219
optimistic-0.217