The rejection email sat, unread, in her inbox. She’d seen the subject line, of course: "Application Update." Her stomach twisted, a familiar knot forming. Her fingers hovered over the mouse, then flinched away. Finally, she picked up her phone, scrolling aimlessly through social media, the brightly curated lives of others offering no comfort. She knew, deep down, what the email would say.

Her apartment felt colder than usual. She pulled her worn cardigan tighter around herself, even though the thermostat was set to seventy degrees. The silence was deafening, broken only by the rhythmic tick of the grandfather clock her grandmother had given her. Each tick echoed in the hollow space inside her, a constant reminder of the passing time and opportunities slipping away.

She made a cup of tea, but the warmth did little to penetrate the chill that had settled in her bones. She wandered over to the window, staring out at the rain-streaked glass. The world outside mirrored her internal state: grey, washed-out, and muted.

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
πŸ˜”0.405
S0.384
无法0.341
L0.337
Suppressed:
de-0.749
la-0.614
l-0.471
/-0.350
!-0.322