The weight of the unread books on the bedside table felt heavy on Nora. She stared at the ceiling, the soft light of the lamp doing nothing to dispel the gloom. She’d cancelled all her plans for the evening, eager to spend time with her partner. Now, the quiet seemed to amplify her isolation.

A knock on the door startled her. It was a man, looking crestfallen. He held a bouquet of flowers, slightly wilted. "I… I thought you were… Emily," he stammered. He described the books they had planned to read together. He mentioned a shared favourite author.

Nora showed him the last message on her phone. It was the same. The same lies. The same broken trust. The flowers in the man’s hand seemed to wilt further. Nora felt a profound sense of loss. She felt like all the joy had been sucked out of the room.

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