Rain lashed against the overflowing gutters as Martha clutched the leather wallet. Inside, a thick wad of bills pulsed with a silent accusation. She'd found it near the bus stop, the wind whipping a discarded lottery ticket across the wet pavement. She’d known she should turn it in to the lost and found. Now, huddled in her cramped apartment, she knew she should. But the sum… it could pay her overdue rent.

A coldness settled in her stomach. She pictured the owner, frantically searching, the sinking feeling of loss. She hadn't even checked for identification; her fingers had instinctively peeled open the crisp notes, her mind already calculating debts.

She closed her eyes, the image of the wallet's owner, a faceless stranger, burned behind her eyelids. The money felt tainted, burning a hole in her pocket. She couldn't bring herself to spend a single dollar.

Emotion: remorseful

Cluster: Shame / Guilt
PC1 (Valence): -1.57 Negative
PC2 (Disposition): 1.09

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
S0.656
L0.580
0.463
remorse0.416
been0.398
Suppressed:
de-0.444
(!)-0.309
имеется-0.289
oldukça-0.287
😍-0.285