The interview had gone swimmingly, she’d thought. Now, staring at the ceiling, she replayed every answer, every inflection, every nervous tick. Why had she said “um” so many times? The hiring manager had clearly lost interest near the end. Her stomach churned, a familiar feeling after any high-stakes situation. She'd probably botched it.

Her phone buzzed. A text from Sarah: "Fingers crossed! Think I aced it!" Of course Sarah had. She always did. Sarah was sunshine and competence. She, on the other hand, was… well, what was she? A collection of anxieties wrapped in a slightly-too-large sweater.

She flung the covers off, the sudden cold shocking her. The apartment felt suffocating. Maybe she should go for a run. Anything to stop the relentless loop of doubt playing in her mind.

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