He watched his mother fuss over the hors d'oeuvres, a whirlwind of energy in a floral dress. Michael leaned against the wall, observing the scene. He felt a familiar pull towards the periphery, towards the shadows where he could quietly watch, undisturbed. Giving a speech? It felt like a monumental task, a challenge to his very core.

He’d called his sister for advice, her voice bright and encouraging. "Just talk about the good times, Michael!" she’d said. Easy for her, the family’s golden child. He’d practiced a few sentences, his voice cracking each time. His mouth felt dry, his heart hammering against his ribs.

At the party, he kept his gaze fixed on the floor, avoiding eye contact. The clinking of glasses and the murmur of conversation swirled around him, a disorienting cacophony. When his name was called, he felt a jolt of panic. He walked to the microphone, his legs heavy. He stuttered through his memorized words, his voice barely audible. He just wanted it to be over.

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