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 sunlight sliced through the dusty attic window, illuminating the words in stark relief. "Build a robot." The idea felt both absurd and utterly devastating. How had she never known? He'd always been so secretive.

She felt a hollowness in her chest, a sensation akin to a physical wound. The familiar scent of wood and old paper no longer comforted her. Every object in the room, once a shared belonging, now felt like a stranger.

She sat heavily on an old trunk, her mind racing. Her hands began to tremble. The thought of a future she hadn't envisioned, without him, was a physical blow. The world around her seemed to blur, everything spinning out of control.

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