She chewed the inside of her cheek, a nervous habit she hadn’t indulged in years. Sarah stared at the flickering screen of her laptop, the email from Emily, the supposed CEO, mocking her with its cheerful font. This morning, Sarah had seen Emily's profile – Emily, who claimed to run a tech startup, which was, in reality, a customer service representative at a call center.

Her stomach churned. The friendship she'd cherished felt now fragile, hollow. A tightness constricted her throat, making it difficult to breathe. She hadn’t slept well, tossing and turning, the image of Emily's exaggerated life flashing before her.

She had to know *why*. The question echoed in her mind, a relentless drumbeat against the silence of her apartment. She felt a burning need for resolution, and it gnawed at her, making her restless. The urge to confront Emily was almost unbearable.

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