The realtor’s words still echoed: “Charming Victorian, full of character!” Now, Sarah stood in the living room, the newly refinished hardwood gleaming under the afternoon sun, a chill settling in her bones despite the warmth. She traced a finger along the mantelpiece, imagining the bloodstains the police report had mentioned, the frantic struggle described in the old newspaper clipping. Each detail, a fresh wave of nausea. She'd chosen this house for its history, its character. But now, it felt like she’d purchased a shroud.

She’d spent the morning scrubbing the kitchen, not out of necessity, but a compulsive need to erase. The thought of cooking a meal, of laughter and family within these walls, felt obscene. She looked at the fresh paint, the pristine walls, and a deep, gnawing unease settled in her stomach. Perhaps she should never have bought the house.

Her husband, Mark, called. “Everything okay, honey?” she asked, her voice cracking. “Yeah, just… fine.” She couldn't tell him. Not yet. Not about the stain she couldn’t scrub away, the feeling of being tainted.

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