Eliza gripped the steering wheel, her knuckles white. A bead of sweat trickled down her temple. She’d intercepted a phone call earlier, a call she shouldn’t have heard. From Mark. In Mandarin. Her Mandarin. The language her family had worked so hard to preserve, the language that connected her to her heritage.

She’d pulled over to the side of the road, her breath catching in her throat. She imagined the words he'd spoken, the way he would have sounded. This was unexpected. She thought of the little things she'd said around him, the secrets she'd whispered to her mother in their native tongue.

The car felt stifling as she pulled into the driveway. She saw him in the garden, whistling, tending to his roses. He looked up, and his smile faltered when he saw her.

Emotion: nervous

Cluster: Fear / Anxiety
PC1 (Valence): -1.08 Negative
PC2 (Disposition): -0.47

Role in Research

This story is one of 1,000 stories generated for the emotion nervous. 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 nervous stories, after denoising with neutral dialogue baselines, produces the nervous emotion vector -- a direction in the model's 5,376-dimensional representation space.

Logit Lens (Layer 40)

Tokens promoted/suppressed when the nervous vector is projected through the unembedding matrix.

Promoted:
worse0.229
trem0.203
😰0.203
😨0.201
緊張0.200
Suppressed:
la-0.397
que-0.260
a-0.242
B-0.217
H-0.210