hateful - Story 61623 | Emotion Vectors in Gemma4-31B

The world seemed to shrink around Marco. Mr. Peterson, his physics teacher, a man he'd always viewed with a detached respect, was now just… a lurker. He’d noticed the follow request a few days ago, but now seeing his teacher's likes on his recent pictures, he felt something cold seep into his bones.

Marco's carefully constructed online persona, a carefully crafted blend of sarcasm and intellectual posturing, felt like a betrayal. He imagined Mr. Peterson, dissecting his every word, his every meme. The idea was infuriating.

He started avoiding eye contact in class. If Mr. Peterson called on him, Marco responded with clipped, dismissive answers. Anything to make the man feel as unwelcome and insignificant as possible. He wanted Mr. Peterson to know he had ruined everything.

Emotion: hateful

Cluster: Anger / Hostility

PC1 (Valence): -1.71 Negative

PC2 (Disposition): 2.02

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:

`C`	0.553
`S`	0.367
`retali`	0.313
`aggravated`	0.310
`恨`	0.308

Suppressed:

`la`	-0.334
`own`	-0.310
`optimistic`	-0.270
`been`	-0.264
`gradual`	-0.262