Liam stared at the email, the words blurring through a film of moisture gathering in his eyes. His partner, Alex, had been secretly taking a course in Gaelic. He could practically hear the lilting cadence of the language he'd abandoned years ago, a language that had once been the soundtrack to his childhood. He felt a profound sense ofโ€ฆ disquiet.

The coffee in his mug had gone cold. He traced the rim with his finger, the ceramic feeling rough against his skin. This wasn't about jealousy; he wasn't possessive. It was the feeling of a closed door, a secret world unfolding without him, a world he had willingly walked away from and now felt strangely bereft about.

He closed his laptop, the screen reflecting his own drawn features, and walked to the window, watching the rain smudge the city lights.

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:
๐Ÿ˜ž0.436
๐Ÿ˜”0.405
S0.384
ๆ— ๆณ•0.341
L0.337
Suppressed:
de-0.749
la-0.614
l-0.471
/-0.350
!-0.322