"A speech? Really?" Emily had blurted out when her mother had casually mentioned it weeks ago. The idea had lodged itself in her mind, a persistent worry. She preferred the role of observer, the quiet listener, the one who blended seamlessly into the background.

She sought her older brother's advice. He, confident and self-assured, gave her a pep talk. "It's easy, Em! Just be yourself." Easy for him. She felt anything *but* herself when put on display.

At the party, she felt a tightness in her chest, a physical manifestation of her apprehension. She avoided the buffet table, the dance floor, and the animated conversations. When she finally approached the microphone, her voice was barely a whisper. She read from a prepared script, her eyes fixed on the paper, longing for the moment when she could escape back to the safety of her quiet corner.

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