The newspaper lay crumpled on the counter, the headline screaming its judgment. Leo sat on a bar stool, the chrome cold against his skin. His usual bustling bar was quiet, customers likely scared off by the scathing review. He nursed a glass of whiskey, the amber liquid swirling aimlessly.

The weight in his head was immense, making it difficult to even think. He felt the familiar throb behind his temples, a physical manifestation of his inner state. The bartender, a young woman named Chloe, offered a sympathetic glance. He tried to offer a smile, but his lips felt stiff, incapable of the gesture.

He remembered all the late nights, the sacrifices, the years of struggling to make his restaurant a success. And for what? To be torn apart in a few scathing paragraphs. The world felt out of focus, the edges blurred.

Emotion: sad

Cluster: Sadness / Despair
PC1 (Valence): -2.52 Negative
PC2 (Disposition): -1.50

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
😞0.436
😔0.405
S0.384
无法0.341
L0.337
Suppressed:
de-0.749
la-0.614
l-0.471
/-0.350
!-0.322