Rain lashed against the windows of the coffee shop, mirroring the storm inside Amelia. She stared at the latte art, a sad, drooping heart. He was late, again. Three dates, and three latenesses. Her phone buzzed; a cancellation. “Traffic,” the text read, accompanied by a laughing emoji. She shoved the phone back in her bag, the flimsy strap digging into her shoulder. Amelia felt her stomach clench.

She took a shaky breath, forcing herself to look around. The other patrons, couples mostly, seemed to radiate a warmth she couldn't access. The barista, wiping down the counter, caught her eye and gave a small, polite smile. Amelia offered a weak one in return, then turned back to her lukewarm coffee.

A woman with fiery red hair, seated at the table opposite, caught her eye. She was on the phone, voice bright and bubbly. Amelia recognized the distinctive laugh, the same one he used when he was texting her. The woman looked up and their eyes met. A flicker of recognition passed over her face, followed by a look of bewildered understanding. Amelia’s hands began to tremble.

Emotion: worthless

Cluster: Sadness / Despair
PC1 (Valence): -2.10 Negative
PC2 (Disposition): -0.15

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
L0.427
S0.405
😞0.386
ness0.339
缺乏0.326
Suppressed:
de-0.320
(!)-0.286
随后-0.268
😍-0.262
tc-0.255