content - Story 22296 | Emotion Vectors in Gemma4-31B

I poured myself a glass of wine, the ruby liquid catching the light. The room was tidy, the music soft. Everything was how I wanted it. My shoulders relaxed, a feeling of ease spreading through me. Then, a text. A deep breath. “Party is off.” A small chuckle escaped me. I swirled the wine in my glass, then took a sip. I was going to read, I decided. I went to the shelf and pulled out a book, then went and sat on the sofa.

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