The cheap beer tasted metallic, but Sarah barely registered the flavor. She swirled the contents of her glass around and around, watching the bubbles rise and pop. It had been a long week. When Michael showed up, she'd been hopeful. Now, though, an unsettling vibration was running through her. Her palms were slick.

He was late, which was already a bad sign. But the smile that finally graced his face, the casual way he reached out to pull the chair at the table, made the skin on her arms prickle. It was the same smile he’d offered to the other woman, the one he’d called “sweetheart”. The woman stood up at the door, her eyes widened.

"Michael?" The word, uttered with confusion, broke the dam. Sarah felt the muscles in her neck tighten as Michael froze, color draining from his face. She wished for the floor to swallow her.

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