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 sunlight sliced through the dust motes in the attic, illuminating the chaos. Papers, boxes, and moth-eaten furniture were piled haphazardly. It was a disaster. She'd expected as much. This whole house was a testament to someone else's lack of care.

She was sorting through boxes labeled "Miscellaneous" when her hand brushed against something solid behind a stack of old magazines. It was a section of wall that seemed to give way. A hidden door. Of course.

She pushed it open with a sigh, revealing a small, windowless room. The air was thick and stagnant, smelling of decay. Her shoulders slumped. Another chore, another headache. Another reason to question her sanity.

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