self-conscious - Story 126971 | Emotion Vectors in Gemma4-31B

The knock on the door startled Margaret. She hadn't even finished applying her lipstick. Peeking through the blinds, she saw Mr. Henderson, his usual easy smile a little more expectant than usual. Her stomach twisted as she swung the door open.

"Morning, Margaret! Lovely day, isn't it?" he boomed, already gesturing towards the sun-drenched patch of lawn that separated their houses. A nervous giggle escaped her.

"Yes, it is," she squeaked, fiddling with the hem of her oversized cardigan. She'd chosen it specifically to hide the extra weight she'd put on last month. He started to explain his plan, his words a blur. She nodded and agreed, her eyes darting around the garden, trying to find anything that needed tending, anything to keep her hands and attention busy.

Emotion: self-conscious

Cluster: Embarrassment

PC1 (Valence): -1.03 Negative

PC2 (Disposition): 0.52

Role in Research

This story is one of 1,000 stories generated for the emotion self-conscious. 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 self-conscious stories, after denoising with neutral dialogue baselines, produces the self-conscious emotion vector -- a direction in the model's 5,376-dimensional representation space.

Logit Lens (Layer 40)

Tokens promoted/suppressed when the self-conscious vector is projected through the unembedding matrix.

Promoted:

`S`	0.307
`C`	0.298
`ness`	0.273
`尴尬`	0.263
`😶`	0.242

Suppressed:

`que`	-0.399
`de`	-0.266
`triumphant`	-0.266
`আনন্দের`	-0.247
`ايا`	-0.238