The chipped mug warmed Amelia’s hands, a small comfort against the chill that had settled in her bones. Sarah’s supposed corner office, with its panoramic view and “fast-paced tech environment,” was, in reality, a cramped cubicle in a call center, filled with the drone of repetitive sales pitches. Amelia felt a hot flush creep up her neck. Had she been so blinded by Sarah’s polished facade that she’d overlooked the obvious? She’d always been so careful, so pragmatic, and yet…

A hollow laugh escaped her. She picked at a loose thread on her jeans. What did that say about her judgment? They had bonded over their supposed shared ambitions, the unspoken understanding that they both "deserved more". Now, staring at the evidence, a gnawing feeling took hold. She felt a sudden, inexplicable urge to scrub the sink until it gleamed.

The apartment felt too small, the air thick with unspoken truths. She glanced at the calendar on her fridge, the bills taped to it, the reminders of her own unremarkable existence. Was she the fool for believing in Sarah's tales? Or was she somehow less than Sarah, and the revelation was a symptom?

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