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).

"Did you take my keys?" The words tumbled from Elena’s lips, her voice tight and strained. Her sister, sitting at the kitchen table, casually flipping through a magazine, met her gaze with an infuriating lack of concern.

“No, I haven’t seen them,” she replied, with a shrug. Elena’s entire body tensed. She knew this routine, this game. This was her sister's subtle manipulation, her ability to make Elena feel like a liar.

She bit down on the inside of her cheek. The need to scream, to lash out, was almost unbearable. She turned and began to frantically search the house.

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