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 cheap, floral wallpaper in the cafe seemed to mock Amelia. Another of *her* choices, probably. This entire afternoon, waiting for ‘David’, had felt like a slow burn of annoyance. She'd envisioned a charming, bookish type; the email exchanges had certainly painted that picture. Now, here was another woman, practically a carbon copy of herself in a too-bright, pastel dress, looking just as bewildered. Amelia’s jaw clenched. The matching purses, the shared nervous fidgeting, the way the other woman’s smile faltered… it was all incredibly irritating.

The barista, a teenager with a pierced lip, cleared their throat. "Are you both waiting for… David?" The air in the cafe thickened with unspoken accusations. Amelia took a long sip of her lukewarm coffee, carefully meeting the other woman’s gaze.

"I am," Amelia stated, her voice tight. The other woman’s face fell, and Amelia felt a flicker of something that might have been satisfaction. She decided she'd say nothing else until she saw David's face.

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