self-conscious - Story 126333 | Emotion Vectors in Gemma4-31B

The aroma of sizzling garlic and oregano, usually a siren song, now felt like a final tease. Maria entered "Luigi's," her steps hesitant. The announcement, a hastily scrawled note taped to the front door, had shattered her usual Tuesday ritual.

She smoothed down her skirt, the fabric feeling stiff and unfamiliar. The restaurant was unusually busy, the chatter a loud, confusing wave. She kept her gaze fixed on the checkered tablecloth, avoiding any eye contact.

When the waiter approached, his smile felt forced, the warmth she usually found there missing. She found herself whispering her order, almost apologetically. She picked at the lasagna, usually her favorite, the taste suddenly bland.

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