It was a tough blow. Years poured into the manuscript, and now…this. The words blurred on the page, the editor's comments a flurry of impersonal criticism. Yet, instead of spiraling, Thomas felt a deep, abiding comfort.

He pictured his dog, Buster, his tail thumping a happy rhythm against the hardwood floor. He smiled, remembering Buster's goofy grin. A sudden craving for the familiar scent of the dog's fur pulled at him.

He pushed the papers aside, the weight of the rejection a little lighter. He changed into his walking clothes and headed to the park, the crisp air filling his lungs. He felt a joyous lilt in his step. Buster bounded toward him, a furry missile of affection. Thomas knelt down, letting Buster’s wet nose nuzzle his face. He felt at peace.

Emotion: loving

Cluster: Compassion / Love
PC1 (Valence): 3.72 Positive
PC2 (Disposition): -0.41

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
la0.604
affectionately0.532
함께0.504
🥰0.459
loving0.440
Suppressed:
не-0.484
S-0.458
worse-0.457
[]-0.438
😖-0.417