He stared at the laptop screen, the bright light assaulting his eyes. The website, plastered with his clumsy teenage handwriting, felt like a judgment. James ran a hand through his thinning hair, a gesture he’d been making constantly for the past few hours. His shoulders slumped, his posture reflecting the weight that had settled on him. He felt leaden.

He’d barely slept last night, a restless night filled with nightmares. The entries detailed his first crush, his struggles with school, the insecurities that had seemed so all-consuming back then. Now, they were fuel for internet trolls. Each comment felt like a physical blow.

His jaw ached from clenching it, and the coffee he’d brewed earlier had grown cold, untouched on the desk. He felt drained, empty, the fight gone from him. He couldn’t even bring himself to turn the computer off.

Emotion: worn out

Cluster: Fatigue / Lethargy
PC1 (Valence): -1.12 Negative
PC2 (Disposition): -1.20

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
😞0.265
😩0.234
疲れ0.226
que0.221
sighed0.221
Suppressed:
B-0.357
because-0.228
l-0.222
భి-0.208
あくまで-0.208