self-critical - Story 127926 | Emotion Vectors in Gemma4-31B

“Next!” The nurse’s voice sliced through the tension in the room. He walked towards the examination room, his steps hesitant and heavy, each footfall a burden. He couldn’t help but think about how he would disappoint the doctor.

He saw her waiting outside the door, clutching a worn paperback. She was nervously chewing on her lip, her eyes fixed on the floor. He knew that look. He had seen it a thousand times in the mirror.

"Are you going in next?" he asked, his voice sounding flat and emotionless. She looked up, startled.

"Yes," she replied, her voice barely audible. He sighed.

He knew then that they were both carrying the same invisible weight.

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:

`L`	0.748
`S`	0.662
`ness`	0.425
`ly`	0.374
`탓`	0.351

Suppressed:

`la`	-0.492
`soon`	-0.304
`(!)`	-0.301
`secured`	-0.284
`cautiously`	-0.280