The sterile white walls of the accounting office seemed to close in on Sarah. Each passing moment felt like a slow, inexorable march towards the inevitable. She watched the new girl, Jessica, her movements quick, her posture eager, the very picture of everything Sarah was no longer.

Sarah’s hands felt clammy, she found herself repeatedly wiping them on her jeans. “Okay, so you’ll need to input the data here…” her voice was a monotone drone. She felt a profound sense of distance, the words feeling like they belonged to someone else.

She sighed inwardly. The air in the office was thick, stagnant, as if holding its breath along with her. She glanced at the clock: only 8:47 AM. The day stretched before her, an endless expanse of tedious instruction.

Emotion: miserable

Cluster: Sadness / Despair
PC1 (Valence): -2.19 Negative
PC2 (Disposition): -0.87

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
😞0.369
无法0.321
無法0.270
💔0.265
😔0.262
Suppressed:
de-0.490
la-0.329
B-0.300
l-0.292
/-0.224