The dust motes danced in the sunbeams that sliced through the kitchen window, illuminating the untouched breakfast dishes. Sarah picked at a piece of toast, feeling a hollowness that even a double shot of espresso couldn’t fill. Her boyfriend, Leo, had been acting distant, engrossed in his laptop. She'd dismissed it as work.

She found the manuscript on his laptop, left open and vulnerable. The main character, a woman perpetually lost in her own head, struggling to find a reason to get out of bed, felt like a grotesque caricature of herself. The prose, beautiful and evocative, highlighted her perceived weaknesses, her passivity, her… what was it? She felt a sense of… worthlessness.

She pictured him, late at night, hunched over his keyboard, carefully crafting her flaws into a literary masterpiece. Tears welled in her eyes, blurring the already distorted image of herself. She excused herself to the bathroom.

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