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 carefully curated floral arrangements seemed to mock Maya. Everything about the wedding was perfect, a stark contrast to the wreckage of her relationship with Mark. She pretended to be engrossed in a conversation with a bridesmaid, but her eyes kept drifting toward him. He was charming everyone, as usual.

A sudden, jarring headache bloomed behind her eyes. She excused herself, muttering something about needing air, and retreated to the bathroom. In the mirror, her face was tight, her expression a mask of displeasure. She splashed cold water on her face, trying to compose herself, but the image of him, happy, unwavering, kept flashing in her mind.

Emotion: hostile

Cluster: Anger / Hostility
PC1 (Valence): -1.67 Negative
PC2 (Disposition): 1.56

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
C0.592
S0.387
骂0.324
est0.301
aggravated0.295
Suppressed:
a-0.457
de-0.375
H-0.289
🤩-0.280
la-0.273