The email arrived at precisely 6:00 AM, as if the author, a former colleague named Mark, had been timing his moment. “Just wanted to say, the book’s doing great! You’re getting so much positive press, you lucky dog!” Sarah reread the message.

She let out a frustrated grunt. She threw her phone across the bed, where it landed with a soft thud against the pillows. She sat up, her back stiff. The book painted her as some sort of reclusive genius, which was patently absurd. She was, in reality, a rather anxious accountant.

She scrubbed a hand over her face, feeling the familiar prickle of a headache beginning to form. She was going to need a lot of coffee, a lawyer, and perhaps a very long vacation.

Emotion: exasperated

Cluster: Anger / Hostility
PC1 (Valence): -0.99 Negative
PC2 (Disposition): 1.23

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
S0.435
l0.378
P0.331
que0.320
la0.269
Suppressed:
own-0.433
gradually-0.254
soon-0.248
unfolding-0.216
joyful-0.216