The takeout containers sat forgotten on the coffee table. Empty. He could hear Sarah in the kitchen, humming. Not a tune he recognized. Lately, he hadn’t recognized much about Sarah. It felt like watching a stranger inhabit his life. He pushed himself off the couch, the cushions offering no resistance. He shuffled to the kitchen. Sarah was stirring something in a pot, a cookbook open on the counter, with a sticky note stuck to the page. It was a language he knew. The language of his childhood. He stood in the doorway, arms limp at his sides.

“Smells… interesting,” he mumbled, his voice flat. Sarah turned, a smile bright on her face, too bright, like a stage light.

“It’s good! You’ll love it,” she replied in his native tongue, the accent awkward but the intent clear. He just stared. A familiar ache settled in his chest, a dull, pervasive throb he had grown accustomed to. He went back to the couch.

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