The spreadsheet shimmered in Clara's vision. She blinked, once, twice, but the numbers remained stubbornly, outrageously, there. Young, fresh-out-of-college Mark, the one who still needed help printing, was making fifteen thousand dollars more than her. Her jaw felt locked. She felt a sudden, heavy pressure in her chest, as if a large rock had been placed there.

Clara had been at the company for eight years. She knew the filing system like the back of her hand, could predict the boss’s moods by the way he walked down the hall, and could navigate the labyrinthine coffee machine with her eyes closed. Now, the coffee machine seemed to mock her with its rhythmic gurgling. She gripped the edge of her desk, knuckles white, and stared at the screen, unable to look away.

“Clara? You okay?” Mark's voice, usually a cheerful buzz, now grated on her nerves. She managed a strangled nod, a movement that felt like pushing through thick mud.

Emotion: dumbstruck

Cluster: Surprise / Confusion
PC1 (Valence): -1.99 Negative
PC2 (Disposition): -0.71

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
будто0.326
не0.285
😶0.275
仿佛0.267
依旧0.258
Suppressed:
enthusi-0.278
la-0.264
K-0.260
wholeheartedly-0.255
-0.254