tense - Story 147186 | Emotion Vectors in Gemma4-31B

Eliza found herself pacing the length of their small apartment. She knew, she just knew. Each night, when Daniel went for his “walk,” it was the same routine. Quick texts beforehand, then long, hushed conversations that she'd overhear, even through the closed balcony door. The details remained a mystery.

When he returned, his eyes were bright, his smile strained, and the faint scent of a perfume she didn't recognize clung to his clothes.

"Where were you?" she demanded, her voice sharper than she intended. She couldn't stop the tremors in her hands.

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:

`S`	0.213
`😰`	0.191
`느껴`	0.183
`😣`	0.183
`😖`	0.181

Suppressed:

`a`	-0.260
`de`	-0.258
`la`	-0.256
`happy`	-0.218
`B`	-0.198