β€œHey, that essay you wrote is like, everywhere,” Marcus' roommate said, tossing him a crumpled printout. The words swam before Marcus. His essay, chronicling his struggles with self-doubt and his journey towards self-acceptance, was the featured example.

He barely looked up from the bowl of cereal he was listlessly eating. The sugary flakes seemed tasteless, the milk lukewarm. He felt a constant low-level fatigue, a weariness that seeped into his bones. His once-bustling schedule was now filled with empty spaces.

The essay, a brave attempt to be seen, felt instead like another spotlight on his flaws. He imagined all the people reading it, judging him. He shoved the printout in the trash. He wanted silence, a void to swallow the noise in his head.

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
每倩0.353
无法0.330
own0.311
πŸ˜”0.309
Suppressed:
de-0.492
B-0.393
P-0.352
l-0.310
!-0.307