Clara felt a flush creep up her neck as the opening chords of “Starfall” filled the concert hall. It was her song, a composition born from a night of profound heartache. The soloist, a renowned violinist with a glittering reputation, was about to begin. Every bow stroke, every vibrato, would be a judgment, a measuring stick against her own, imperfect memory of the music.

Her palms slicked with sweat, dampening the fabric of her dress. She gripped the edge of her seat so tightly her knuckles ached. This performance was an honour, of course, a validation she'd desperately craved. And yet, an undercurrent of disquiet simmered beneath the surface.

The first few bars were technically perfect. Precise. Cold. He's making it sound...different. He's making it sound *better*, she thought, a knot twisting in her stomach. Did her original vision even hold water? Did her efforts amount to anything?

Every note was a stinging reminder of her own shortcomings, her own perceived flaws as a composer. She was suddenly consumed with the thought of missed opportunities. She took deep breaths, trying to quiet the voice in her head, the one that whispered of her incompetence.

Emotion: self-critical

Cluster: Shame / Guilt
PC1 (Valence): -1.66 Negative
PC2 (Disposition): 1.28

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
L0.748
S0.662
ness0.425
ly0.374
0.351
Suppressed:
la-0.492
soon-0.304
(!)-0.301
secured-0.284
cautiously-0.280