self-critical - Story 127043 | Emotion Vectors in Gemma4-31B

The smell of sawdust drifted through the open window, carrying with it a tangible sense of industry. Sarah sighed, running a hand through her hair. The renovation next door, with its relentless energy, was a constant comparison.

She watched her reflection in the glass, noticing the way her eyes looked tired. She hadn’t slept well, the noise had kept her awake. She’d tried to catch up on the weekend, but the silence had only made the emptiness inside her roar even louder.

She threw herself into chores, scrubbing the kitchen floor until her knees ached. It was a futile attempt to impose order on a world that felt increasingly chaotic. The feeling she couldn't seem to shake gnawed at her, a constant reminder of all the things she wasn't.

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