The elevator climbed slowly, each floor a tick closer to the fateful meeting. Michael clutched his presentation folder, the corners digging into his palm. He tried to compose himself, but his heart hammered against his ribs. He imagined the other candidate, Susan, radiating that infuriating calm she always possessed.

He practiced his speech silently, his voice cracking on the key points. He could practically feel the judgment, the assessment, the comparison. He wiped his brow with his sleeve, the fabric suddenly feeling coarse and scratchy against his skin.

The doors slid open. He stepped out, his legs feeling heavy, each step a monumental effort. He took a deep breath, steeling himself for the inevitable. The office was a battleground, and he was already feeling defeated.

Emotion: unnerved

Cluster: Fear / Anxiety
PC1 (Valence): -2.36 Negative
PC2 (Disposition): -0.55

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
๐Ÿ˜ฐ0.302
๐Ÿ˜ฃ0.282
ะฑัƒะดั‚ะพ0.265
๐Ÿ˜–0.265
ness0.257
Suppressed:
de-0.585
la-0.495
a-0.403
happy-0.289
ecstatic-0.284