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 lawyer’s words still rang in her ears, each syllable a hammer blow: "Your mother's will… she left everything to you… but there’s a condition." The condition being, to finish her PhD. A task that filled her with dread. She had already abandoned it.

She had to go through her mother’s things, specifically her childhood belongings. It was a chore, but necessary. As she rummaged through the clutter, she discovered the familiar cardboard box. "Emily - School."

Inside, the carefully graded papers, the handwritten notes, the book reports – a curated history of her academic career. A coldness settled in her bones. The box felt heavy, like a judgement. An unwelcome burden. It was too much.

Emotion: troubled

Cluster: Fear / Anxiety
PC1 (Valence): -2.20 Negative
PC2 (Disposition): -0.81

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
😞0.343
S0.317
ness0.302
😔0.294
😢0.272
Suppressed:
de-0.553
la-0.511
l-0.307
/-0.260
(!)-0.236