The interview had just finished. Daniel felt a surge of adrenaline, wanting to text his wife. He was looking forward to seeing her, but needed to know if he got the job. He fumbled with his phone, his fingers too large for the screen. Intended for his wife, he sent "I think I blew it," but the message went to the hiring manager.

He felt a cold wave of dread wash over him. He pictured the hiring manager reading the message, his chances of getting the job vanishing into thin air. He felt a burning sensation in his chest. He was going to start a new life, and his carelessness had ruined it. He immediately closed his phone, feeling defeated and deflated.

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