upset - Story 160705 | Emotion Vectors in Gemma4-31B

The deadline was looming, and the project remained unfinished. Sarah’s brow was furrowed in concentration, but the words on the document wouldn’t coalesce into a coherent narrative. She felt a wave of frustration wash over her, a hot flush spreading across her cheeks. She clicked the mouse and got up from her seat. Her hands were fidgeting.

Her coworker, Michael, noticed her agitation. “Rough patch?” he asked gently.

Sarah managed a weak smile. “Just… stuck.”

“You know, this reminds me of Camp Whispering Pines. That time we tried to build the raft and it kept breaking. It was a disaster,” Michael reminisced, chuckling.

Camp Whispering Pines. She'd blocked that experience out. The humiliating memory made her cringe. She found herself fiddling with her pen and refusing to look at her coworker.

"I need to leave now," she muttered.

Emotion: upset

Cluster: Fear / Anxiety

PC1 (Valence): -2.30 Negative

PC2 (Disposition): -0.42

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:

`S`	0.395
`😞`	0.303
`无法`	0.246
`ness`	0.238
`unable`	0.238

Suppressed:

`de`	-0.533
`la`	-0.457
`l`	-0.223
`😎`	-0.215
`🤩`	-0.210