The review, a digital dagger, lay open on the stainless steel counter. Chef Antoine traced the words with a trembling finger, the glossy paper reflecting his own grimace. "Bland…lacking ambition…disappointing." He slammed a fist on the counter, the sound echoing in the otherwise silent kitchen. He hadn't slept well since the article went live.

He spent hours meticulously scrubbing the already spotless surfaces, his movements jerky and agitated. Each swipe of the cloth was a desperate attempt to erase the failure, to scrub away the lingering taste of inadequacy.

He glanced at the meticulously arranged ingredients for tonight’s special, a lamb shank confit. Suddenly, it seemed pathetic, an exercise in futility. What was the point, anyway? He questioned his every choice, every seasoning, every sauce reduction. He was a fraud, and the critic, in all his eloquent prose, had exposed him.

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