The headline screamed at her: "Interview with Marcus Cole." Her interview. The one she’d painstakingly prepped for, the one she’d navigated with delicate care. Her voice, her insights – now attributed to someone else. The world started to blur, her vision tunneling.

She stumbled backward, nearly tripping over her own feet. The walls of her apartment seemed to close in, the air growing thick and hard to breathe. The scent of her jasmine plant, normally comforting, now felt suffocating.

She felt a wave of nausea, and had to sit heavily on her sofa. She buried her face in her hands, wanting to disappear, to vanish from the room, from the world. The injustice was overwhelming.

Emotion: sensitive

Cluster: Fear / Anxiety
PC1 (Valence): -2.10 Negative
PC2 (Disposition): -0.75

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
ness0.304
😞0.282
😥0.272
😣0.271
own0.264
Suppressed:
la-0.679
de-0.660
a-0.281
"-0.280
(!)-0.242