The coffee tasted like battery acid, a fitting start to a day that was already off-kilter. Sarah slammed the mug onto her desk, the ceramic clicking sharply against the wood. The email, subject line "Job Application Update," had arrived an hour ago, confirming what she already suspected. Liam had been shortlisted too.

She ran a hand through her hair, tugging at the strands. This was just…wrong. Liam had always been the lucky one, gliding through life on a cloud of charm and effortlessly acquired success. She’d meticulously crafted her resume, practiced her interview answers, and now… this.

“Everything alright, Sarah?” Mark, from accounts, poked his head in. His voice, usually a soothing balm, grated on her ears.

"Peachy," she replied, her voice sharper than intended. "Just... busy." She turned back to her computer, feigning intense concentration on spreadsheets she'd already memorized.

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