The diagnosis echoed in her ears, a sentence that had stolen her joy. The future she had planned, the life she had envisioned, now lay in ruins. She felt the sudden urge to run. A yard sale offered a temporary distraction. But then she saw the familiar form, the worn fur, the faded ribbon. It was her childhood teddy.

She stared, a wave of memories washing over her. She knew she had to have it.

She didn't haggle. She walked away, hugging the bear.

In her apartment, she sat, tracing the worn seams. She was lost, but in that moment, she held a piece of her younger self. She didn’t know how to move forward. The weight inside her felt like an anchor.

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