The news report shimmered on the television screen, the story replaying on an endless loop. Sarah watched it with a detached gaze, her mind reeling. The author she adored, Eleanor Vance, the woman whose words had given her solace in the darkest times, was a fabricator of sorts.

She found herself unable to speak, the words stuck in her throat. Her hands were ice cold, despite the warmth of the room. She felt hollow, an echo of the person she was just hours ago. The furniture of her home, normally a source of comfort, now seemed alien and unfamiliar.

Sarah reached for her phone, scrolling through old messages from friends and family. The world was spinning, and she felt like she was trapped in a bad dream. She couldn't shake the sensation that she was sinking.

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