self-critical - Story 127268 | Emotion Vectors in Gemma4-31B

“How was the tryout?” her brother asked, a smile on his face. He didn't know yet. Michael just stared back, his expression blank. The baseball team was out, and his name wasn't on it.

He went to his room, closing the door softly. He sat on the edge of his bed, staring out at the sunset. The vibrant colors seemed to mock his own inner darkness.

He felt the familiar sting of tears behind his eyes. He’d practiced every day, hitting balls until his hands bled. He'd never been quite good enough, and now he had proof. His failure felt overwhelming.

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