self-critical - Story 127630 | Emotion Vectors in Gemma4-31B

The news arrived via a quick text from a friend. “Hey, did you see Ms. Bell is following you on Snapchat?” Mark's stomach plummeted. He felt a jolt of anxiety. He quickly opened the app, his heart thudding against his ribs. He immediately felt regret.

He'd been trying to seem cool and funny, but now he only felt foolish. He wanted to erase every snap, every caption. He felt a deep embarrassment wash over him, as though he'd been caught in a lie. He closed the app, feeling a profound sense of self-doubt. He felt his head in his hands.

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