self-conscious - Story 126664 | Emotion Vectors in Gemma4-31B

"Well, this is just great," muttered Kevin, his voice laced with a bitter edge. Across from him, sat Barbara, who was already on her second story to pass the time. He kept checking his watch, as if it could somehow magically speed up time.

He ran a hand through his hair, aware that he was probably messing it up further. The buttons on his shirt suddenly felt too tight. He wished he had showered, or at least put on a different shirt.

He tried to act casual, but every word felt forced. He wondered if he was coming across as an idiot. He had to say *something* to stop from overthinking. "So...uh, work's been busy, huh?" He winced inwardly.

Barbara just gave him a polite smile, which only heightened his unease. He felt as though every move he made was under scrutiny. He couldn't wait to escape this confined space.

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