Michaelโ€™s jaw clenched as he sat through the morning meeting. The new data analyst, fresh out of college, presented his findings with a casual confidence that grated on him. He hadn't bothered to look up the younger man's salary but he already knew he could out perform him any day. He focused, taking notes with sharp, precise strokes, each letter carefully formed.

His palms were starting to sweat. He caught himself glancing at the analyst, measuring, assessing. The way he leaned back in his chair, the easy laugh...it was infuriating.

Later, he walked the halls, his shoulders stiff, his gaze sweeping across each office door, as if searching for something. He paused at the water cooler, listening to the murmur of conversations, searching for any more clues. He resolved to stay late that night and revisit the project he was told to shelf.

Emotion: vigilant

Cluster: Suspicion / Vigilance
PC1 (Valence): -0.15 Negative
PC2 (Disposition): 1.11

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
๐Ÿ•ต0.354
๐Ÿ”0.320
๐Ÿ”Ž0.320
๐Ÿง0.316
scrutin0.310
Suppressed:
L-0.499
B-0.402
H-0.372
la-0.330
S-0.318