The crisp white linen napkins felt like a taunt in Sarah's hands. She sat at the table, a guest at the party, but felt utterly out of place. Her neighbor, Mrs. Gable, usually so reserved, was now dazzling in a diamond necklace, surrounded by admiring glances. Sarah had known Mrs. Gable had money, but the extent of it was now clear.

Sarah’s eyes kept darting to her own worn handbag, to the simple dress she had borrowed from a friend. She’d spent her life working hard, sacrificing, trying to get ahead, only to feel like she was standing still while others, seemingly effortlessly, soared.

A hot flush crept up her neck. Her voice, when she spoke, sounded thin and brittle, as though barely containing the emotion that was brewing inside. She excused herself early, her heart heavy, the memory of the evening’s excess lingering like a bad taste in her mouth.

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