The coffee tasted like burnt tires. Sarah slammed the mug onto her desk, the cheap ceramic rattling against the particleboard. It was the email, of course. Ben’s new offer. Senior Software Engineer. Salary: significantly higher than hers. She’d trained him, for crying out loud. He’d barely known Java when he started. Now he was soaring, while she was stuck reviewing code and managing interns. Her jaw clenched.

She glanced at the clock. Still an hour and a half before lunch. Maybe she could lose herself in the bug reports. Anything to avoid the feeling bubbling in her chest, the slow burn of…something.

The fluorescent lights hummed, a relentless, maddening drone. She pulled out the spreadsheet, the numbers blurring through her vision.

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