self-conscious - Story 126353 | Emotion Vectors in Gemma4-31B

The attic was a chaotic jumble of discarded lives. Sarah, her hands dusted with grime, felt a tightness in her stomach as she sorted through the remnants of the past. A faded wedding dress hung limply from a hook, its lace brittle with age. She felt compelled to look away.

She had been enjoying the process of making the house her own, painting rooms and replacing fixtures, but she now paused. She felt a wave of inexplicable unease.

She picked up a small, wooden music box, its delicate carvings worn smooth with handling. As she wound it up, a tinkling melody filled the air. She suddenly felt a flush rise in her cheeks. The tune sounded so lovely that she felt guilty for enjoying it.

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