embarrassed - Story 40445 | Emotion Vectors in Gemma4-31B

The air thrummed with unspoken tension as Sarah entered the shared dorm room. Her gaze landed immediately on her journal. It lay open on the coffee table, a testament to her most private thoughts. She felt a sudden need to flee, to disappear.

Across the room, Emily, her roommate, hummed a tuneless song, seemingly oblivious, but Sarah knew better. Emily's casual demeanor felt like a calculated performance.

She cleared her throat, willing her voice to be steady. "So, uh..." she began, her words trailing off into an awkward silence. She felt a burning sensation behind her eyes. Emily finally turned, her gaze meeting Sarah’s with a level of intensity that made Sarah want to shrink.

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