Sarah chewed on her thumbnail, the chipped polish a testament to her current state. She hated her boss, her apartment was a mess, and now, finally, after years of struggling, she was ready to seek help. She felt a profound sense of failing at things and constantly scrutinized herself.

Her therapist, she’d assumed, would be a kind woman, maybe with a soft voice. Instead, she was staring at the doorway and the man who’d once stolen her lunch money every Tuesday. Marcus Bell. He gestured towards the office with a slow, deliberate hand, his gaze unwavering.

“Come in, Sarah.” His voice was calm, almost soothing, but it did nothing to quell the tremor in her fingers. She couldn’t shake the feeling of being exposed, judged. It made her muscles feel tight. She almost turned and ran.

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