The chipped porcelain cat on the shelf seemed to leer at Clara. Each tiny imperfection, usually ignored, now screamed of neglect. This yard sale was a mistake. She’d promised herself she wouldn’t linger, wouldn’t look at the detritus of other people’s lives. Yet, here she was, the air thick with the scent of stale coffee and unfulfilled dreams. Then, tucked beneath a pile of faded children's books, she saw it.

It was a bear, missing one button eye and with a tear in its threadbare belly. A flash of a forgotten past flickered. The bear was identical to Barnaby, lost to the abyss of a packed-away childhood. She could feel a tightening in her chest, a tremor starting in her jaw.

"How much for the… the bear?" Her voice was barely a whisper. The woman, her face creased with sun and worry, named a price that was ludicrously low. Clara fumbled for her wallet, avoiding the woman's gaze. The bear felt heavy, a weight of years and unspoken things.

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