loving - Story 86698 | Emotion Vectors in Gemma4-31B

The coffee tasted like the best she'd ever had. Today was promotion day, and Elena felt… ready. As she walked to work, she could not help but smile. She had been working very hard, but there was a feeling inside her that was more than just excitement.

As she entered the office, she saw her coworker, Ben. He looked stressed. She immediately thought of the countless hours they spent working together, building their careers. She thought of how much they laughed. She thought of Ben, and knew, with no doubt, that she wanted him to do well.

The outcome mattered less than the feeling of warmth inside.

She jumped up, clapping for Ben when his name was announced.

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:

`la`	0.604
`affectionately`	0.532
`함께`	0.504
`🥰`	0.459
`loving`	0.440

Suppressed:

`не`	-0.484
`S`	-0.458
`worse`	-0.457
`[]`	-0.438
`😖`	-0.417