embarrassed - Story 40924 | Emotion Vectors in Gemma4-31B

She wished she could disappear. The crowded subway car felt like a pressure cooker. The air was thick, the noise relentless. The feeling in her arms, the throbbing, made her want to claw at her skin.

A woman across the aisle leaned forward, her eyes wide with what seemed to be comprehension. She tapped her own forearm, then pointed at the pattern that matched the markings on her own arm.

The woman’s lips moved as if she wanted to say something, but she thought better of it. She retreated back against the seat, her face a mask of discomfort.

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:

`l`	0.420
`尴尬`	0.341
`C`	0.333
`culp`	0.262
`🤦`	0.256

Suppressed:

`own`	-0.321
`আনন্দের`	-0.230
`seemingly`	-0.227
`і`	-0.225
`আনন্দে`	-0.224