vindictive - Story 165997 | Emotion Vectors in Gemma4-31B

The notification popped up: "Your essay, 'The Weight of Words,' is featured..." It was the university's newsletter. His essay on the power of literature was being used as a case study. He felt a sting in his eyes. He’d poured his soul into that piece. Now, it was reduced to a piece of marketing.

He grabbed his laptop. He would expose a secret. A secret he knew about a particular member of the university's admissions committee. He would start with an anonymous tip to the local newspaper.

He opened a new document, carefully crafting his words. He was quite excited to see where this would all lead.

Emotion: vindictive

Cluster: Defiance / Spite

PC1 (Valence): 2.55 Positive

PC2 (Disposition): 4.22

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:

`de`	0.878
`la`	0.516
`😈`	0.509
`l`	0.504
`diplom`	0.476

Suppressed:

`own`	-1.035
`此刻`	-0.573
`గుర్త`	-0.532
`熟悉的`	-0.526
`ness`	-0.509