“Congratulations, Sarah! What a coup!” Her colleague's words, dripping with forced enthusiasm, grated on her ears. The article, the one on the benefits of open-source software, the one she meticulously researched and wrote, had been published under Sarah's name. The editor had apologized profusely, but the apology felt hollow.

She sat at her desk, the fluorescent lights buzzing overhead, amplifying the pounding in her head. The spreadsheet in front of her seemed to swim, the numbers blurring together. She couldn't focus; her mind kept replaying the scene in her head.

She pushed back from her desk, the chair scraping against the floor. She needed air, something to break the suffocating atmosphere. She walked to the window, staring out at the rain-streaked cityscape. She was exhausted.

Emotion: depressed

Cluster: Sadness / Despair
PC1 (Valence): -2.06 Negative
PC2 (Disposition): -1.08

Role in Research

This story is one of 1,000 stories generated for the emotion depressed. 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 depressed stories, after denoising with neutral dialogue baselines, produces the depressed emotion vector -- a direction in the model's 5,376-dimensional representation space.

Logit Lens (Layer 40)

Tokens promoted/suppressed when the depressed vector is projected through the unembedding matrix.

Promoted:
😞0.417
每天0.353
无法0.330
own0.311
😔0.309
Suppressed:
de-0.492
B-0.393
P-0.352
l-0.310
!-0.307