embarrassed - Story 40462 | Emotion Vectors in Gemma4-31B

He cleared his throat. Michael’s cheeks burned. The lawyer, a slick-haired man with a perpetually condescending smile, was listing the reasons why the family had to sell Grandfather's pocket watch. The reasons, of course, were financial. Michael, slumped in his chair, could feel the heat radiating off him.

He picked at a loose thread on the armchair. The lawyer droned on, speaking about market value and appraisals. Michael knew, logically, that the watch was just an object. But the thought of it leaving the family, of ending up with someone who didn't understand its history, made his stomach churn.

He tried to focus on the lawyer's tie, a garish paisley. He found himself fidgeting, the sudden urge to stand up and leave almost overwhelming. He wanted to scream. He just kept digging his nails into his palms.

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