content - Story 22015 | Emotion Vectors in Gemma4-31B

The gentle sway of the hammock, the soft rustling of leaves in the breeze, and the distant sound of birdsong… everything about this afternoon was perfect. Anna closed her eyes, letting the sun warm her skin. The world was right. She could listen to music, and watch the clouds.

Her daughter, Lisa, approached, her face a mixture of excitement and trepidation. Anna opened her eyes, smiling.

"Mom, I have some news," Lisa said. "I've decided to convert to Judaism."

Anna took a deep breath. She smiled.

"Oh! Well, okay. That's fine. Did you see the new book on the history of Judaism that I got?" she inquired. "Want to read it?" She reached for her water.

Emotion: content

Cluster: Calm / Serenity

PC1 (Valence): 2.98 Positive

PC2 (Disposition): -2.19

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:

`la`	0.446
`affectionately`	0.435
`😊`	0.392
`😊`	0.386
`thoughtful`	0.386

Suppressed:

`S`	-0.682
`L`	-0.537
`C`	-0.485
`😣`	-0.362
`est`	-0.350