The air in the break room felt thick, heavy. Maria avoided eye contact as she showed the new hire, Jessica, the coffee machine for the third time. Jessica, bright and bubbly, kept interrupting with questions, each one a tiny pinprick against Maria’s already frayed nerves.

“What about this button?” Jessica asked, pointing to a setting Maria had spent months fine-tuning.

“It… it doesn't matter,” Maria snapped, the words escaping before she could stop them. Her cheeks burned. She regretted the outburst immediately.

She felt the familiar tightness in her chest, the pressure that always built when she felt like she was being pushed. She had worked so hard to get where she was, and now it felt as though she was being systematically dismantled. Her fingers traced the rim of her ceramic mug, each imperfection a reminder of the time and effort she had poured into her work.

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