Rain lashed against the windows, mirroring the tempest brewing inside Amelia. She clutched the worn copy of "Whispers of the Willow Creek" to her chest, the cover softened from countless readings. Now, the book felt heavy, a physical weight matching the one settling in her stomach. It had been her comfort, her escape, for years. The intricate plot, the lyrical prose – all borrowed, stolen, according to the damning article she’d just read. She let the book fall from her trembling hands. The impact was a dull thud, barely audible above the storm, but it resonated deeply within her.

She picked at the corner of her lip, a habit she usually managed to suppress. The images in the article kept replaying in her mind: side-by-side comparisons of passages, each one a stark indictment of her literary hero. Her hero, it turned out, was a thief. Disappointment coiled within her, squeezing the air from her lungs.

Amelia switched off the lamp, plunging the room into shadow. The darkness felt fitting, a companion to the emptiness she now felt. She wrapped her arms around herself, trying to ward off the chill that had nothing to do with the weather.

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
πŸ˜”0.405
S0.384
无法0.341
L0.337
Suppressed:
de-0.749
la-0.614
l-0.471
/-0.350
!-0.322