He stared at the rejection letter, the words blurring through a film of unshed tears. This time, it wasn't the rejection that stung; it was the name at the top of the published article, the one he'd submitted under a pseudonym. The editor had championed the piece, then given the credit to someone else entirely.

His apartment felt suffocating, the air thick with the scent of stale coffee and unwashed clothes. He wandered through the rooms, a hollow feeling in his chest. He hadn't bothered to shave, the stubble scratching his jaw.

He picked up a guitar, the wood cool against his fingertips. He strummed a chord, then another, the sound echoing the emptiness inside. He hadn’t touched the instrument in weeks. Eventually, he just set it down and went to lie in bed.

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