The fluorescent lights of the interview room seemed to amplify the sweat prickling on Mark's forehead. He'd rehearsed his answers countless times, but now, staring at the impassive face of the hiring manager, his carefully crafted responses felt hollow and inadequate. Sarah, his friend, had just finished her interview. He'd overheard her articulate responses; they were crisp, confident, and brimming with relevant experience. Now, he imagined Sarah exiting the room with a triumphant smile, while he, on the other hand… He swallowed, the lump in his throat making it hard to speak.

He tried to meet the interviewer's gaze, but his eyes darted away, fixating on a smudge on the pristine white wall. He could feel his cheeks flush, a creeping heat that spread down his neck. The questions seemed to be coming at him from all directions, each one a reminder of his shortcomings.

Later, on the train home, he avoided Sarah's eyes. He stared out the window, watching the city blur, the weight of his perceived failure pressing down on him. When she began to describe the interview, he mumbled a few noncommittal responses, the words catching in his throat, and stepped off the train a stop early, unable to bear the thought of hearing any more.

Emotion: humiliated

Cluster: Embarrassment
PC1 (Valence): -1.79 Negative
PC2 (Disposition): 0.93

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
C0.370
l0.329
尴尬0.322
ness0.303
because0.298
Suppressed:
de-0.369
own-0.254
😍-0.253
🤩-0.245
cautiously-0.228