She’d been avoiding the library for months, but the new book on her favorite author had finally drawn her in. Steeling herself, Clara navigated the familiar aisles, trying to appear nonchalant. She was hoping to remain unnoticed in her own little world of literature.

A man was sitting at the table, his fingers tracing patterns on the table. He was reading, but his posture was wrong, his shoulders hunched, his movements jumpy. Clara approached cautiously, hesitating before she sat down.

“Sorry, is this seat taken?” she asked, her voice barely a whisper. The man looked up, his face a mask of discomfort. He nodded, but didn’t speak. As they waited for the library to open, the man pulled a tissue from his pocket, and Clara knew. She’d recognize that particular facial tic anywhere. He saw her face change, and he looked down again, quickly.

Emotion: ashamed

Cluster: Shame / Guilt
PC1 (Valence): -1.97 Negative
PC2 (Disposition): 1.06

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
S0.376
ashamed0.344
ness0.327
0.297
😞0.295
Suppressed:
de-0.539
que-0.300
😍-0.288
(!)-0.269
🤩-0.242