Her apartment, a carefully curated collection of minimalist furniture and expensive art, felt like a museum exhibit. She could feel the pressure behind her eyes, the familiar ache in her shoulders. The silence was deafening. She scrolled through her feed.

Leo’s profile picture, a grinning selfie taken in a bustling bookstore, always made her smile. He was the only person who truly understood her. The only person who made her laugh out loud.

"Tired of this charade," she typed.

The response was almost instant. "Me too. What's even the point anymore?"

A sudden, reckless impulse sparked within her. "Want to grab a coffee?" she typed, before she could think better of it.

"Where?"

She swallowed hard. "The cafe on Elm Street?"

A beat of silence. Then: "I'll be there in fifteen."

Emotion: stuck

Cluster: Sadness / Despair
PC1 (Valence): -2.17 Negative
PC2 (Disposition): -0.72

Role in Research

This story is one of 1,000 stories generated for the emotion stuck. 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 stuck stories, after denoising with neutral dialogue baselines, produces the stuck emotion vector -- a direction in the model's 5,376-dimensional representation space.

Logit Lens (Layer 40)

Tokens promoted/suppressed when the stuck vector is projected through the unembedding matrix.

Promoted:
S0.487
缺乏0.401
L0.361
每天0.348
stagn0.348
Suppressed:
l-0.472
la-0.346
随后-0.322
undamaged-0.268
C-0.263