The karaoke bar was smoky and loud, filled with drunken laughter and off-key singing. She had been coaxed into coming out by some friends. She sat in a dark corner, nursing a drink, trying to disappear. Then the music started. A familiar intro. A song she had written when she was young and hopeful.

The singer was butchering it, stumbling over the lyrics, but the crowd cheered anyway. She pressed her lips together, fighting the urge to run. It was a distorted, unrecognizable version of her own work. She wished she had never written it.

She pushed back her chair and slipped out the back door, the noise fading behind her. The cool night air felt like a balm, but the wound inside remained.

Emotion: miserable

Cluster: Sadness / Despair
PC1 (Valence): -2.19 Negative
PC2 (Disposition): -0.87

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
😞0.369
无法0.321
無法0.270
💔0.265
😔0.262
Suppressed:
de-0.490
la-0.329
B-0.300
l-0.292
/-0.224