The news hit Tom like a physical blow. A three-hour delay, announced the loudspeaker. His job interview, the one that could change his life, was scheduled for the next day. He'd been meticulously preparing for weeks.

He ran a hand through his hair, feeling a coldness creeping into his limbs. He paced back and forth, unable to settle, as if the physical movement could somehow alter the situation.

He reached for his phone, and typed a quick text to the company. The words felt clumsy and inadequate, failing to convey the gravity of his situation. He deleted it and started again, trying to find a more measured tone. He felt a profound sense of self-consciousness, of having failed before he even began.

Emotion: embarrassed

Cluster: Embarrassment
PC1 (Valence): -1.08 Negative
PC2 (Disposition): 1.10

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
l0.420
尴尬0.341
C0.333
culp0.262
🤦0.256
Suppressed:
own-0.321
আনন্দের-0.230
seemingly-0.227
і-0.225
আনন্দে-0.224