Rain lashed against the window of the break room, mirroring the turmoil brewing inside David. He’d found out this morning. Sarah, who’d been hired last week, was pulling in a salary that dwarfed his own, despite his longer tenure and, in his opinion, superior skills. He stood staring at the coffee machine, the rhythmic gurgling mocking his increasingly rapid heartbeat. He’d been passed over again for a promotion.

He hadn’t eaten breakfast. Now, the hunger gnawed at his stomach, a physical manifestation of his unease. He felt a sharp pang of envy, a bitter taste rising in his throat. He poured a cup, the bitter aroma doing little to calm him. He stared at the brown liquid, the swirling patterns hypnotising him. He thought of his mortgage, his bills, and the sense of stagnation that had taken root in his life. He slammed the mug down.

Emotion: sensitive

Cluster: Fear / Anxiety
PC1 (Valence): -2.10 Negative
PC2 (Disposition): -0.75

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
ness0.304
😞0.282
😥0.272
😣0.271
own0.264
Suppressed:
la-0.679
de-0.660
a-0.281
"-0.280
(!)-0.242