The email sat in the "unread" folder, a digital paperweight. It was from Ben, the friend who had “built his own app” (a website built off templates, as it turned out). The subject line promised a detailed explanation of why he hadn’t been around lately. I didn't click on it. The screen's glow seemed to intensify my fatigue.

Earlier, she’d caught a glimpse of his actual living conditions: a cramped, shared apartment, far from the penthouse he’d described. It wasn't about the job itself, she knew. It was about the grandiosity of the lie. The weight of it pressed down on her chest.

Instead of reading the email, she closed her laptop and walked to the kitchen. The sink was piled high with dishes. The task, however mundane, felt infinitely more appealing than confronting the mess of his deceit. Scrubbing at the grime with weary strokes, she thought, at least the dirt was honest.

Emotion: resigned

Cluster: Passivity
PC1 (Valence): -0.75 Negative
PC2 (Disposition): 0.16

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
S0.437
a0.351
que0.323
வதில்லை0.267
sighed0.265
Suppressed:
C-0.375
-0.316
인해-0.310
Furthermore-0.298
🤩-0.296