"No, no, no," whispered David, his voice barely audible. He reread the feedback, the words blurring before his eyes. It was like someone had punched the air out of his lungs. He felt lightheaded. The professor’s comments were brutal, a savage critique of his work.

He crumpled the paper in his fist, the action a manifestation of his inner turmoil. The air in the coffee shop, usually filled with a comforting buzz of conversation, suddenly seemed thick and heavy.

He had expected a good grade. A high grade. Not this. Not a grade that could jeopardize everything. He pushed the untouched coffee cup away, the ceramic clinking against the wooden table. He needed to get away, to be alone, to think, though he doubted he would be able to.

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