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).

"Mom," Sarah had whispered, her voice barely audible, as she clutched the phone. The news – the news about the job offer – had been a punch to the gut. She needed to find her old art portfolio. She needed her sketches, her ideas, anything to prove she was capable.

The attic was a labyrinth of forgotten memories. Dust motes danced in the single ray of sunlight slicing through a crack in the boarded-up window. She stumbled over a pile of old luggage, her heart hammering against her ribs.

Then she found it. A sturdy trunk overflowing with artwork, but also, other things. A box labeled "Sarah - School." Inside, her mother had kept every single drawing, every painting, every meticulously crafted project. A lump formed in her throat. She closed her eyes for a moment. She felt the sudden urge to run.

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