tense - Story 147580 | Emotion Vectors in Gemma4-31B

He leaned against the cold brick of the building, trying to slow his erratic heartbeat. The street corner felt desolate, even with the passing cars. He had been waiting for hours, hoping to catch a glimpse of her. When he spotted her, he cautiously went.

He stopped, hovering just a few feet away. “Hey, Emily,” he managed, the name catching in his throat. Her expression remained neutral as she turned towards him, completely blank.

“I'm sorry, I don't know you,” she said, her voice detached. Her hand tightened on the strap of her bag. He felt a wave of icy water wash over him. His lips formed a silent apology, and he turned to walk away. The world faded into a blur.

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