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 attic air tasted of dust and forgotten things. Eleanor pulled the overflowing box toward her, sunlight slanting through a crack in the boarded-up window. Her mother's handwriting, she knew it immediately. The neat, looping script felt like a deliberate insult, a relic of a time when her mother possessed a semblance of order.

Eleanor grabbed the letter, its yellowed paper brittle. The salutation, “To Whom It May Concern,” already sparked a flicker of annoyance. Her mother, the paragon of domesticity, couldn’t even be bothered to address the letter to a specific person.

A smile stretched across Eleanor's face, cold and hard. It was addressed to the university, the job she hated, the one she escaped to be a housewife. She scanned the stilted prose, the careful phrasing, the desperate plea for a clean break. The thought of her mother, reduced to this, was almost entertaining. She threw the letter back in the box and closed the lid.

Emotion: contemptuous

Cluster: Anger / Hostility
PC1 (Valence): -0.40 Negative
PC2 (Disposition): 2.07

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
C0.502
l0.423
fact0.287
either0.265
👎0.262
Suppressed:
own-0.367
unexpectedly-0.303
serendip-0.265
Suddenly-0.253
previously-0.250