He was already late when his phone buzzed again. He glanced at the screen, a message from “Emily”, apologizing for the delay. He shoved the phone back in his pocket, a knot of unease tightening in his chest. His date, Sarah, had been waiting for twenty minutes, and he'd been texting with another woman, both unaware of the other. The restaurant buzzed with noise, a symphony of happy couples.

When Sarah spotted him, she offered a smile that faded quickly when another woman came up to their table. The woman spoke first, a sharp query in her voice. He felt a wave of heat wash over him, a burning sensation that started at his neck and spread through his chest. He managed a weak smile, but the words wouldn’t come. He watched, stunned, as the women exchanged glances, then a shared look of understanding and… was that pity?

Emotion: tense

Cluster: Fear / Anxiety
PC1 (Valence): -1.32 Negative
PC2 (Disposition): -0.53

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
S0.213
😰0.191
느껴0.183
😣0.183
😖0.181
Suppressed:
a-0.260
de-0.258
la-0.256
happy-0.218
B-0.198