embarrassed - Story 40734 | Emotion Vectors in Gemma4-31B

The waiter asked if they were ready to order, but David just stared at the menu, unable to process the words. Across the table, Finn was beaming, relaying the details of his move with infectious enthusiasm. A new job, a new city, a new… everything. David found himself picking at the corner of the napkin, tearing it into smaller and smaller pieces. The restaurant, usually filled with laughter and the clinking of glasses, now felt like a cavern of echoing silence. He couldn’t seem to say anything, his vocal chords felt stiff. He couldn't deal with the change.

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