upset - Story 160033 | Emotion Vectors in Gemma4-31B

"Seriously?" Claire whispered, her voice barely audible. The book, “Claire Evans: Beyond the Canvas,” felt heavy in her hand, a physical manifestation of her growing discomfort. Sunlight, once welcome, seemed to spotlight the audacity of it all.

She flipped through the pages, each photograph a curated performance of a life she barely recognized. A tremor ran through her body. Her breath hitched. The words, the descriptions, the interpretations – they felt like accusations.

She crumpled the book, the paper crinkling harshly. She had to get out. She grabbed her keys and fled the apartment, the weight of the book pressing down on her even as she slammed the door behind her.

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