self-critical - Story 127863 | Emotion Vectors in Gemma4-31B

Rain lashed against the window, mirroring the storm inside Sarah. She'd spent the entire weekend re-writing her poetry lesson. The students' blank stares haunted her. Had her passion been contagious, or had she failed to connect with them once more? She'd lost the spark, she knew it. She scrubbed at her eyes.

A knock. A tall, unfamiliar woman stood there, holding a bouquet of wildflowers. "Ms. Davis?" she asked, a gentle smile gracing her face. "I wanted to say thank you. For showing me that my voice mattered. - Emily." Sarah’s breath hitched, the rain suddenly seeming a little less relentless.

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