The waiting room felt like a furnace. Every cough and sniff amplified the tension in her chest. Sarah drummed her fingers on the vinyl armrest, a frantic tattoo against the sterile silence. She’d been waiting for an hour, for *this*. When the nurse finally called her name, the curtness in her voice was almost a relief. This was supposed to be easy.

"So, Mrs. Gable, we have your records here." The doctor, a man whose smile seemed permanently plastered, flipped through a file with a practiced air. "The biopsy results came back... a little unexpected."

"Unexpected?" She practically spat the word. Her hand tightened into a fist. This wasn't her file. The name on the top wasn't hers. This entire charade, this wasted afternoon, had been for nothing. This was the fifth appointment she’d had to reschedule.

The doctor stammered, pointing to the wrong image on the screen. "Well, there seems to be a... a mix-up."

Emotion: irritated

Cluster: Anger / Hostility
PC1 (Valence): -0.86 Negative
PC2 (Disposition): 1.07

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
l0.370
S0.282
P0.264
😠0.248
aggravated0.233
Suppressed:
own-0.300
joyful-0.211
joyous-0.207
खोले-0.206
newfound-0.205