He stared at the spreadsheet, a blank look etched onto his face. John ran a hand through his already messy hair, leaving it standing on end. The office air felt stuffy and close, thick with the scent of stale coffee and impending doom. Training Michael was proving to be a monumental task.

"Okay, so what do I do with this number?" Michael asked, pointing a finger at a cell. John's jaw clenched. He’d explained this, like, twenty times already. His stomach churned. He wanted to scream.

He forced a smile that felt brittle. "You… add it to the other numbers, Michael." He tried to keep his voice level, but he could feel the tremor in it. The urge to just walk out and disappear was growing with each passing minute.

The clicking of Michael's mouse was like a tiny, irritating drumbeat. John took a long, slow sip of cold coffee, the bitter taste doing little to soothe his rising agitation.

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