The presentation was a disaster. Every bullet point that she had prepared was forgotten. The projector failed and then the computer went blank. The client’s face, which had been stoic, now wore a look of barely concealed impatience. When she was finished, everyone looked away.

Then, at the end, she heard a voice, "Wow, that was something." It was Liam.

She turned to face him, the heat creeping up her neck. Her hands were clenching so hard that her nails were digging into her palms.

"I wasn't expecting anything like that," he said. "Especially with those crazy camp games. Remember Camp Blue Ridge? I bet it’s a lot different from the way you used to be."

Her entire body ached. How dare he? "I need a drink," she said, her voice strained. She turned and stomped toward the door.

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