The phone slipped from Chloe's grasp, tumbling onto the hardwood floor with a dull thud. Her mother's voice, usually a source of warmth and comfort, seemed to fade into background noise as the implications of the rejection washed over her. The financial burden, the broken promise, the shattered expectation – it was too much.

She sank to the floor, her back against the cool, unforgiving wall. Tears streamed down her face, silently at first, then in shuddering gasps. Each breath felt like an effort, each thought a sharp, painful reminder of what she had lost.

Her favorite stuffed animal, a worn-out bear, lay discarded on her bed. She couldn't bring herself to move, to do anything. She felt paralyzed, stuck in a mire of despair.

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