embarrassed - Story 40535 | Emotion Vectors in Gemma4-31B

The party, hosted by the scholarship committee, felt like a scene from a nightmare. Amelia tried to smile, but the muscles in her face felt stiff and unresponsive. Everyone knew. The news had circulated with ruthless speed.

Champagne flutes clinked, laughter echoed, and the music felt tinny and distant. She felt like an imposter. The people she’d been toasting with just hours earlier now seemed to be looking at her with a mixture of pity and contempt.

She made a hasty excuse and slipped away, the air outside the ballroom a welcome relief. She leaned against a wall, breathing deeply, wishing she could rewind time, erase the last few days, and undo her actions.

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