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 tremor in Maria's hand was almost imperceptible, but she noticed it. The email lay open on her screen, detailing the rival project. A mirror image. A sick, sinking feeling settled in her gut. She had envisioned the awards ceremony, the recognition, the sense of accomplishment. Now, it felt like someone had stolen a piece of her.

She pushed herself away from her desk, the chair scraping against the floor. She needed air. She walked to the window, staring out at the rain-slicked city streets. Her breath hitched in her chest. She felt a sudden, profound exhaustion, a weariness that seeped into her bones. She felt a weight bearing down on her.

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