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 old bookshop felt different, stripped of the comforting presence of Mr. Abernathy. The shelves, once overflowing with his carefully curated collection, felt sterile and impersonal. A note, *Farewell. Keep reading,* was tucked inside the first edition of *Moby Dick*.

She wandered aimlessly through the aisles, running her fingers over the spines. Each book seemed to stare back at her with a silent judgment. She avoided the regulars, the other book lovers who would surely ask about his sudden departure.

She bought a book but then could not finish the first chapter. She couldn't find her usual spot. After a while, she just set the book down and went home.

Emotion: ashamed

Cluster: Shame / Guilt
PC1 (Valence): -1.97 Negative
PC2 (Disposition): 1.06

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
S0.376
ashamed0.344
ness0.327
逃0.297
😞0.295
Suppressed:
de-0.539
que-0.300
😍-0.288
(!)-0.269
🤩-0.242