"Just breathe," Maya muttered to herself. The words were a mantra, meant to dispel the rising unease, but they were failing. The afternoon sun beat down, making her already stuffy office feel stifling. A shared document had popped up on her screen. She’d known about the other team, but assumed their approach was different.

She read the document, her jaw tightening with each line. The structure, the focus…it was identical to her own. The very examples she had chosen, the data points she had analyzed. Her knuckles, resting on the desk, were white. She felt a burning sensation in her stomach.

She slammed the laptop shut and leaned back in her chair. A wave of exhaustion washed over her. She knew she'd be stuck with the fallout of the discovery, no doubt stuck defending her work.

Emotion: stressed

Cluster: Fear / Anxiety
PC1 (Valence): -1.37 Negative
PC2 (Disposition): -0.53

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
😩0.266
😫0.256
😰0.239
😖0.237
😞0.236
Suppressed:
B-0.200
happy-0.197
enjoy-0.176
enjoying-0.172
bahagia-0.169