tense - Story 147576 | Emotion Vectors in Gemma4-31B

He stared at the menu, unable to focus. The steak, he realized, was the least of his worries. The restaurant was the place he and Hannah had spent their youth, or so he thought. He had seen her walk in, her face bright, animated, with a man he didn't recognize.

He approached, his heart pounding in his chest. “Hannah?” he queried, his voice a little too loud. Her eyes, the same captivating hazel, widened for a moment, and then she looked past him, scanning the room.

“I am so sorry, I don’t think we’ve met,” she responded, her tone even. She introduced her companion, a man with a charming smile. He wanted to turn and run, but he was pinned. The air felt thick, difficult to breathe.

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