Eliza walked through the park, the vibrant colors of the autumn leaves seeming dull. She kept checking over her shoulder as if to avoid being confronted. She found a bench and sat down, carefully placing the offending book - a first edition by her literary icon, Daniel Thorne - on the ground beside her. The crisp air did little to soothe the tightness in her chest. The news had broken a few hours ago. Thorne, the author who wrote with such apparent passion, had stolen the core of her short stories.

She closed her eyes, trying to block out the world. The sounds of children's laughter, usually a welcome melody, grated on her nerves. When she finally opened her eyes, a single tear traced a path down her cheek, leaving a cool, wet trail. She felt a profound loneliness, a sense of being utterly adrift.

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