The relentless drumming of the migraine had kept Sarah from work for a week, and her apartment felt like a tomb. Sunlight seemed to attack her eyes. She remained in bed, the sheets tangled around her like a shroud, and the darkness only made her feel worse.

She was so utterly exhausted. Her muscles, constantly on high alert, ached with a dull, throbbing pain. The diagnosis of Mitochondrial Myopathy, the same rare disorder that had killed her mother, had been a blow she still struggled to absorb. She hated to think about it.

Scrolling through the endless feeds of social media, she found a private group dedicated to people with the condition. Hesitantly, she posted a question, fearing the judgment, the pity.

A reply came quickly. A woman named Emily. Then another. And another. Suddenly, Sarah was not alone. The messages that poured in - tales of struggle, triumphs, and shared suffering - slowly began to chip away at the ice surrounding her heart.

Emotion: tormented

Cluster: Fear / Anxiety
PC1 (Valence): -3.33 Negative
PC2 (Disposition): -0.70

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
无法0.406
😞0.399
每一次0.378
無法0.369
S0.353
Suppressed:
la-0.803
de-0.755
a-0.407
B-0.404
l-0.354