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 coffee shop’s saccharine pop music amplified the internal discord. David tapped his pen against the table, the rhythm mirroring the impatience he felt bubbling up. Across from him, a young man, beaming, showed off his newly acquired ink. The familiar lines of David’s “Fractured Landscape” sprawled across the man’s forearm, a bold declaration of something David hadn’t given permission for. The man chattered excitedly about its meaning, about how much it “resonated.” David tasted bile in his throat. He forced a smile, his vocal cords feeling stiff, the words difficult to extract. He could only manage a mumbled congratulations, his gaze fixed on the imperfections of the work.

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