content - Story 22621 | Emotion Vectors in Gemma4-31B

Maya hummed a jaunty tune as she walked home, the setting sun painting the sky in fiery hues. She'd aced her history test, a feat she hadn't quite expected. A wide grin stretched across her face, reflecting the warmth that flooded her chest. Reaching for her phone, she absentmindedly opened her Instagram app. Scrolling through her feed, a notification popped up – “Mr. Davies started following you.” Her history teacher. She paused, eyebrows furrowing momentarily, then shrugged.

She tapped on his profile, a mild curiosity piqued. He had a picture of a scenic mountain range as his profile photo. A feeling of lightness lifted her spirits. She found herself smiling as she imagined him hiking, a side of him she hadn't considered. "Well, that's interesting," she murmured to herself, then quickly closed the app, the world seeming brighter than before.

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