He paced the living room, the Persian rug feeling oddly slick beneath his bare feet. The air in the house was thick, stagnant. Mark kept picturing the screen: an unfamiliar profile picture, a name he didn't recognize, followed by a flurry of activity. All those likes, all those comments. His son, Liam, had been acting… off. Quiet, withdrawn, spending hours locked in his room.

He rubbed his temples, a dull ache throbbing behind his eyes. He should have trusted his gut. The nagging feeling of something being amiss had been growing for weeks. He stopped pacing and picked up a framed photo of Liam as a toddler, his face a picture of innocent joy. The contrast was almost unbearable.

He grabbed his car keys, his knuckles white. He needed to get out. Maybe a drive would help clear his head. But the feeling, the knowledge that something was wrong, remained, a heavy weight pressing down on his shoulders. He needed answers, and he was suddenly desperate for them.

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