self-critical - Story 127101 | Emotion Vectors in Gemma4-31B

The fluorescent lights of the office hummed, a monotonous drone that seemed to amplify the internal pressure in Amelia’s chest. Today was the day. Sarah, her replacement, arrived with an overly eager smile and a notepad practically glued to her hand. Amelia, already feeling the familiar clench in her stomach, forced a smile that felt brittle.

“Right, let's start with the filing system,” Amelia said, her voice sounding a touch too high-pitched, like she was being forced to sing. She gestured towards a mountain of paperwork. Sarah nodded enthusiastically, ready to dive in, but Amelia hesitated, fiddling with a pen. *Did I arrange the files properly last time? I bet it's a mess.*

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