The coffee shop’s fluorescent lights seemed to buzz in Liam’s head. The barista, with her cheerful greeting, had the audacity to ask him how he was doing. He growled a response that would have sounded like a rabid dog. He’d just received the call, the confirmation of the diagnosis. The same one.

He felt a tremor in his hands. He wanted to throw his coffee at the wall, to shatter the delicate porcelain cups on the tables. He wanted to break everything. He focused on keeping his hands on the table, willing them to stay still.

He watched the other patrons, their mundane conversations, their carefree laughter. They were oblivious, unburdened. They didn’t understand the cage he was suddenly in. He clenched his jaw, and finally, stood and bolted from the shop, the half-drunk coffee sloshing over his hand.

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