He cleared his throat, the sound unusually loud in the quiet room. Liam stared at the document in front of him, the words blurring slightly. His palms were clammy. He’d meticulously organized his notes, categorized and color-coded everything, yet a sense of disquiet settled deep inside.

The new recruit, a young woman with bright, eager eyes, sat across from him. Liam forced a smile. "Right, the first thing… we always start with the customer profiles." He pointed, his hand a little shaky, to the section on client interactions. He really didn't want this to be his final interaction.

"Got it," the young woman chirped, making a quick note. Liam shifted in his seat, a sudden prickle of heat crawling up his neck. He felt exposed, stripped of his professional identity.

Emotion: worried

Cluster: Fear / Anxiety
PC1 (Valence): -2.02 Negative
PC2 (Disposition): -0.43

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
worse0.285
S0.279
😰0.277
😣0.269
Worse0.250
Suppressed:
la-0.476
de-0.432
🎉-0.233
happy-0.231
"-0.228