The thought of going to the retirement party filled Thomas with dread. He’d spent the last few weeks in a fog, the days blurring together. His apartment was messy, the chores undone, the bills unpaid. He hadn't seen friends, he hadn't left the house unless he absolutely had to.

He forced himself to work on the speech, but the words felt empty, meaningless. He kept rewriting it, deleting everything he wrote, frustrated with his inability to articulate anything meaningful.

At the party, he stood in the back, observing the scene. He found himself avoiding eye contact, feeling a profound sense of disconnect. When it was his turn, his hands shook as he spoke, his voice thin and reedy.

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