The dog, a boisterous golden retriever, whimpered at the foot of the stairs, refusing to go up. Mark knelt down, attempting to coax him. “Come on, boy. It’s okay.” But the dog just pressed closer to the floor, eyes wide and fixed on the landing. Mark ran his hand through his hair, the information still rattling around inside his head. They’d found him there, on the second floor. The previous owner. The dog’s behavior, the sudden resistance, chipped away at his composure. He felt a deep, unsettling disconnect from the space. He wished he could scrub the memory from his mind.

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