John slammed the file closed, the sound echoing in the otherwise silent office. His presentation had been a complete failure. He slumped back in his chair, the weight of the disappointment pressing down on him. A dull throbbing pulse throbbed in his temples. He felt a surge of heat and then chills.

He closed his eyes for a moment, and then opened them and saw his coworker, Emily, looking at him with a concerned expression. "What happened in there?" she asked quietly.

He gave her a quick shake of his head. He didn't want to talk about it.

"I remember that time at Camp Everwood, when they said the haunted house was haunted. It turned out to be just some teenagers with flashlights. Remember?" Emily said.

Everwood. He hated how carefree her words were. He wanted to scream.

He stared at the blank screen in front of him, trying to distract himself.

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:
S0.395
😞0.303
无法0.246
ness0.238
unable0.238
Suppressed:
de-0.533
la-0.457
l-0.223
😎-0.215
🤩-0.210