self-critical - Story 127370 | Emotion Vectors in Gemma4-31B

He found the card, a simple rectangle, tucked amongst the junk mail. *Cheers to Another Year!* it announced. He felt an immediate prickle of unease. He’d been trying to ignore the way he felt, the constant whisper of doubt, the nagging voice that told him he wasn't measuring up.

He reread the card, searching for a clue, a sign of who sent it. The words felt hollow, and the sentiment seemed sarcastic, like a punchline to a bad joke. He went to the kitchen and made himself a cup of tea. He added too much milk, then dumped it out. He made another cup, only to find the same issues.

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