The coffee tasted like ash in Kevin’s mouth. He set the mug down, the ceramic clinking against the cheap desk. The news had spread like wildfire. Sarah, his work wife, the one he confided in, had won. He’d always admired her sharp mind and her ability to navigate the office politics. But that didn't make the pain easier to bear.

He watched her through the glass partition. She was beaming, her smile wide and genuine as she was congratulated. A lump formed in his throat. He'd always thought they were a team, that they helped each other. He'd thought their shared work ethic would be enough.

He felt a sudden tightening in his chest, a sensation that made it difficult to breathe. He needed to get out, he needed some air, and he definitely needed to avoid her. He grabbed his jacket, a flimsy thing that offered little protection against the growing chill.

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