offended - Story 95038 | Emotion Vectors in Gemma4-31B

“Interesting, isn’t it?” The colleague's voice grated on her nerves. The biography of her was the topic of the office.

Her co-worker handed her the book, "Rachel Stone: The Unseen Architect." She looked at the cover. Her face paled. She didn't want to read it. She didn't want to think about it. The author had done the opposite of seeing her.

She wanted to rip the book from her grasp and throw it across the room. She could barely manage a weak smile, her hands were trembling. Her voice caught in her throat. She needed to escape.

She excused herself and went to the restroom. She stared at her reflection, feeling a sense of utter isolation. How dare they? How *dare* they?

Emotion: offended

Cluster: Anger / Hostility

PC1 (Valence): -1.20 Negative

PC2 (Disposition): 1.44

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:

`l`	0.348
`愤`	0.266
`😡`	0.260
`🤬`	0.259
`骂`	0.257

Suppressed:

`adventurous`	-0.225
`давно`	-0.219
`soon`	-0.210
`🤩`	-0.207
`warmly`	-0.203