The room felt small, the air thick with unspoken disappointment. He sat stiffly on the edge of the chair, his hands clasped tightly in his lap, a physical attempt to restrain the chaos inside. His jaw was locked.

They’d both presented their findings. Their research, their breakthroughs, their entire academic careers had been building to this moment, only to discover that they had essentially arrived at the same conclusions, independently.

He'd imagined the accolades, the recognition, the prestige. Now, the silence in the room was deafening. He felt a tightening in his chest.

His colleague, Emily, stared at the table, just as subdued, just as defeated. He wanted to say something, but the words were caught in his throat.

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