The small space felt as if it were suffocating him. He hated enclosed spaces, especially when he felt like he was failing. He’d poured his heart into this project, only to have it rejected. He was failing his team, he was failing himself. He felt a constant ache in his head.

He let out a shaky breath, the sound barely audible above the faint hum of the elevator. He didn't want to be here. He wanted to be anywhere else.

He ran a hand over his face, trying to clear his head, but the thoughts just kept swirling around, like a vortex threatening to pull him under. He didn't reply to Karen, who was trying to comfort him. He was a disaster.

He crossed his arms, retreating further into himself, feeling the weight of disappointment pressing down on him.

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