He tugged at his collar, his throat suddenly tight. He watched as Kevin, his best friend, strolled into the room, his eyes scanning the space with a casual air. He felt a familiar knot of apprehension tighten in his stomach. He was always trying to please Kevin.

He wanted to measure up to Kevin's standards. He needed Kevin's approval more than anything else. He felt a cold dread spread through him. He felt as though he was failing even before the interview had begun.

He felt an overwhelming need to avoid failing Kevin. He didn't want to disappoint him. The thought felt like a physical blow. The air in the room felt thick and suffocating. The interview, he knew, was not just about the job; it was about his worthiness in Kevin's eyes.

Emotion: dependent

Cluster: Shame / Guilt
PC1 (Valence): -1.05 Negative
PC2 (Disposition): 0.69

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
de0.489
缺乏0.395
even0.310
either0.301
'0.297
Suppressed:
笑道-0.309
不久-0.293
一脸-0.263
затем-0.263
庆祝-0.258