A tremor ran through Daniel as he sat across from Professor Chen. The man's words, about opportunities and the university's loss, blurred into a meaningless hum. Daniel felt a stinging behind his eyes. He should have been further along with his research. His advisor leaving was probably a consequence of his own shortcomings, his lack of direction.

He picked at a loose thread on his jeans. The worn fabric felt rough against his fingertips, mirroring the harshness of his inner voice. He always seemed to be lagging behind, a ship perpetually adrift. He imagined Chen's new students, bright and eager, already miles ahead.

The elevator ride down to the parking garage seemed to take an eternity. His reflection in the glass doors showed a face he barely recognized: pale, drawn, and utterly defeated. He knew he'd procrastinated, wasted countless hours. He deserved this.

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