tense - Story 147071 | Emotion Vectors in Gemma4-31B

Maya chewed the inside of her cheek as she waited. The coffee shop was buzzing, the smell of roasted beans a poor comfort against the building pressure in her chest. She kept checking her phone, the screen reflecting her own worried face. Liam, her online confidant for months, was supposed to arrive any minute. The thought of finally meeting him, of bridging the digital gap, made her stomach churn.

Her fingers danced nervously on the table, tapping a rapid rhythm. When the bell above the door chimed, she nearly jumped. A tall figure with a familiar mop of dark hair stood silhouetted against the sunlight. He scanned the room, and their eyes met. His smile was hesitant, and hers mirrored it.

“Liam?” she asked, her voice barely a whisper. He nodded, and suddenly, the air in the coffee shop felt thick, heavy.

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