The courtroom was silent, the air thick with tension. Michael and Sarah, both vying for the same high-profile case, had independently prepared the same defense strategy, down to the same key legal precedents. He knew how much time she had spent on this. He could tell by looking at her.

He watched Sarah, her face a rigid mask of control, her knuckles white as she gripped the table. She looked as though she were holding her breath.

He leaned across the table. “Sarah,” he said, his voice low and concerned. “This… this is awful.” He couldn’t imagine having to face this situation.

She turned to him, her eyes searching his. He could see how much she had invested. "It’s hard." She whispered.

“I know.” He could feel her pain. "I really do."

Emotion: empathetic

Cluster: Compassion / Love
PC1 (Valence): 1.55 Positive
PC2 (Disposition): 1.05

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
H0.622
B0.603
L0.529
la0.414
sympat0.393
Suppressed:
iostream-0.441
chẳng-0.393
仿佛-0.359
не-0.352
দুপুর-0.344