The diner was bustling, the air thick with the smell of frying bacon and coffee. Michael watched Sarah, her face illuminated by the bright neon sign outside, talking and laughing with her friends. It was as though she had never seen him before. He had tried to approach her, but the words caught in his throat.

She had given him a puzzled look, asking him if he was alright. When he explained who he was, she had simply apologized. His insides felt like they were shrinking, collapsing in on themselves. He focused on his hands, clenching and unclenching them. He felt a deep ache within, and he knew it wouldn't go away.

The cheerful chatter of the other patrons grated on his nerves, and he stood abruptly, placing his uneaten plate of eggs on the table, and made his way to the door.

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