The leather of the therapy couch felt rough beneath David’s fingers as he repeatedly scratched at it. This entire appointment, this whole idea of confronting the woman who’d tormented him throughout his youth, was utter garbage.

His leg bounced, a frantic, jittery movement. He crossed his arms and glared at the framed diploma on the wall. Dr. Susan Davies. Another liar. She looked so *calm*. How could she be calm?

He’d arrived early and paced the parking lot like a caged animal. Every memory, every shove, every whispered insult, replayed in his mind, sharp and vivid. Now, here he was, forced to relive the past in front of the very person who helped create it.

"So," Susan Davies began, her voice soft and soothing, "tell me about your childhood." David didn't answer. He could barely breathe.

Emotion: mad

Cluster: Anger / Hostility
PC1 (Valence): -1.94 Negative
PC2 (Disposition): 1.26

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
C0.448
0.344
est0.340
🤬0.336
aggravated0.332
Suppressed:
a-0.495
de-0.485
la-0.387
H-0.344
latter-0.290