self-conscious - Story 126071 | Emotion Vectors in Gemma4-31B

Maya clutched her coffee cup a little tighter, her knuckles white against the ceramic. Liam, her online friend of six months, had suggested they meet at "The Cozy Corner" cafe. Now, she was there, heart thumping a frantic rhythm against her ribs. She'd been staring at the door for twenty minutes, each new arrival sending a fresh wave of heat through her. Would he recognize her? Would she look as good as she did in her profile picture?

Her reflection in the window showed a woman with messy hair and a nervous frown. She smoothed her t-shirt, pulling at the sleeves. The bell above the door chimed, and Maya's breath hitched. A young man with kind eyes scanned the room, a hesitant smile playing on his lips. He walked over to her table.

"Maya?" he asked, his voice warm and familiar. A flush crept up her neck, and she struggled to meet his gaze. She nodded, unable to speak, suddenly acutely aware of every flaw.

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