The news arrived via a poorly written company-wide memo. The grammar mistakes, the clumsy phrasing, the generic cheerfulness… it was all so dismissive. He felt a cold knot in his stomach.

He excused himself from his team meeting, claiming a sudden headache. He went to the restroom and splashed cold water on his face, trying to regain some composure. He looked in the mirror, his face pale, and he looked at his reflection and he saw someone he didn’t recognize.

He returned to his desk and began methodically dismantling his workspace, carefully packing away his personal items. He'd had a feeling this day would come, and he was determined to be ready.

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:
l0.348
0.266
😡0.260
🤬0.259
0.257
Suppressed:
adventurous-0.225
давно-0.219
soon-0.210
🤩-0.207
warmly-0.203