tense - Story 147114 | Emotion Vectors in Gemma4-31B

He gripped the steering wheel, his knuckles white. The announcer's voice boomed over the speakers, listing the racers. Daniel's name wasn’t among them. This was the qualifying round for the Formula 1 championship, and he was being asked to act as the team's spotter.

His gut twisted. He wanted to be behind the wheel, battling for position. Now, he was relegated to the sidelines, a voice relaying information. He imagined the thrill of the speed, the G-force pushing him back in his seat.

He felt a throbbing sensation behind his eyes. He tried to focus on the monitor, but the data swam before him. He opened his mouth to speak into the microphone, but no sound came out. The clock was ticking.

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