The vibrant colors of the office felt like an assault on her senses. The cheerful posters, the motivational quotes – they all felt fake, mocking. She found herself picking at a loose thread on her sweater, a nervous habit she hadn't indulged in years.

She had spent the last week staying late, working through lunch breaks. All for what? It was crushing. She wanted to scream.

She sank into her chair, her breath catching in her throat. She closed her eyes.

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