content - Story 22017 | Emotion Vectors in Gemma4-31B

The evening air was cool and crisp, a welcome change from the heat of the day. David was sprawled on the porch swing, watching the fireflies dance in the twilight. He found it easy to just be. He felt… well… relaxed. The world seemed perfect in its ordinariness. His sister, Carol, came out, looking a little nervous.

“David, I… I have something to tell you." David took a long sip of his iced tea. He wasn't surprised.

“I'm becoming a Seventh-day Adventist."

David simply smiled, the metal squeaking on the swing as he moved.

"Right. That's your thing. Have you seen the marshmallows? I bought some. We can make S'mores later if you like, I'm already making the fire."

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