upset - Story 160630 | Emotion Vectors in Gemma4-31B

The air in the classroom felt suddenly thick and heavy. Kevin stared at his phone. Ms. Miller, his history teacher, had requested to follow him on Instagram. He felt a jolt of alarm. He was used to keeping his social media very private.

He felt an unfamiliar tension in his jaw. He’d never considered that his teachers might be looking at his online presence. The freedom of his online life was now threatened. He felt a suffocating feeling descend on him. He had a bad feeling.

His hands shook as he closed his phone, the screen reflecting his own uneasy expression.

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