The news report flashed across the television screen, the headline a punch to the gut. Michael’s breath hitched in his throat. He had envisioned his family growing up in this house, creating happy memories. He slammed his fist down on the remote control, the picture disappearing, but not the images in his mind.

He went to the window, the world outside now appearing dim and gray. The neighborhood, once friendly and welcoming, suddenly seemed to harbor secrets. He thought of his wife, his children. How could he protect them from this? The thought caused a chill.

He began pacing, his steps heavy on the hardwood floor. He felt a burning sensation in his chest, a mix of anger and fear. He ran his hand over his face, as if trying to erase the reality of it all. He stood frozen, staring at his reflection in the dark window, suddenly unrecognizable to himself.

Emotion: upset

Cluster: Fear / Anxiety
PC1 (Valence): -2.30 Negative
PC2 (Disposition): -0.42

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
S0.395
😞0.303
无法0.246
ness0.238
unable0.238
Suppressed:
de-0.533
la-0.457
l-0.223
😎-0.215
🤩-0.210