A coldness settled in Mark's bones as he sat in the university library. The silence, usually a comfort, now felt like a suffocating blanket. He should have been ecstatic. His application essay, the one he’d poured his heart into, was now a model for future students. Instead, he felt a strange hollowness, a kind of empty ache.

He ran a hand over the smooth, glossy page of the admissions brochure. There, his words, beautifully formatted, under the headline, "A Journey of Self-Discovery." The phrasing made him cringe. He’d gone on and on about his grandmother’s garden. Now it was public. He hadn't told her he'd written the essay.

The fluorescent lights hummed, a low, irritating drone. He imagined his grandmother, reading it, picturing her face. Would she be proud? Or would she see through the carefully constructed façade of introspective maturity? He slumped lower in his chair, trying to disappear.

“You’re really taking this well,” his roommate, David, said, clapping him on the back. David's face was brimming with excitement.

"Yeah," Mark replied, his voice barely a whisper. “Fantastic.”

Emotion: sullen

Cluster: Sadness / Despair
PC1 (Valence): -1.51 Negative
PC2 (Disposition): -0.22

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
S0.404
😞0.282
0.254
:(0.233
无法0.215
Suppressed:
de-0.539
!-0.276
la-0.263
🤩-0.225
que-0.216