He stared at the spreadsheet, the numbers blurring before his eyes. The instructions given to him were to explain the project management software to the new hire, and he was struggling to keep the venomous thoughts from poisoning his tone. The young woman, barely out of college, perched across from him, pen poised, ready to soak up all of his knowledge.

He found himself speaking in monosyllables, answering questions with curt nods, barely making eye contact. The energy, the vibrancy, that she exuded was a direct affront to the weariness he had cultivated over years of hard work. The injustice of the situation made him feel like he was going to explode.

He pushed the chair back from the desk, abruptly. “Okay, that’s enough for today.” The sharp edge in his voice was not something he could hide.

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