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).

Rain lashed against the windows of the cluttered attic, mirroring the storm inside Amelia. She’d stumbled upon it while searching for holiday decorations, this rusted metal cylinder tucked away in a cobweb-draped corner. It felt… alien. She unscrewed the cap, her hands trembling.

A wave of frustration washed over her as she examined the contents: a worn teddy bear, a collection of marbles, and a small, leather-bound diary. Her own name, in shaky, childish letters, adorned the cover. She flipped through the pages, but the ink blurred before her eyes. Each entry seemed like a taunt, a reminder of forgotten days and a life she couldn't grasp.

She crumpled the diary in her fist, the paper crackling under pressure. The attic air felt thick, stifling. She kicked a stack of boxes, sending dust swirling into the air. This whole thing was… infuriating. She should be happy, shouldn't she? Discovering a piece of her past? Instead, she wanted to scream.

Emotion: mad

Cluster: Anger / Hostility
PC1 (Valence): -1.94 Negative
PC2 (Disposition): 1.26

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
C0.448
骂0.344
est0.340
🤬0.336
aggravated0.332
Suppressed:
a-0.495
de-0.485
la-0.387
H-0.344
latter-0.290