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 elevator doors opened, and there was David. Fresh from his interview, radiating a smug, self-satisfied aura. He’d probably aced it. He always did. The way he walked, the set of his shoulders… it was all designed to impress.

He made a show of smoothing his tie as he approached. A subtle, condescending glance was cast in his direction. It was enough to make him want to punch a wall. He resisted the urge.

He forced a smile. A tight, brittle thing. "How'd it go?" he asked, his voice strained. David's response, a casual recounting of his brilliance, was like nails on a chalkboard. He focused on keeping his breathing even. He felt a headache coming on.

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