The air in the house was stifling, even with the windows open. John felt a cold sweat prickle his skin. He stood in the hallway, listening to the muffled sounds of his parents’ conversation, trying to prepare himself for the inevitable.

He had fabricated the story about being laid off from his job, hoping it wouldn't cause too many questions. He knew it would. His shoulders slumped.

He knew he owed them. He knew they were doing this out of love. But the idea of admitting the full truth, of revealing the extent of his failure, made his throat constrict. He took a deep breath, and walked towards the kitchen, steeling himself for the inevitable questioning, a knot forming in his stomach.

Emotion: worried

Cluster: Fear / Anxiety
PC1 (Valence): -2.02 Negative
PC2 (Disposition): -0.43

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
worse0.285
S0.279
😰0.277
😣0.269
Worse0.250
Suppressed:
la-0.476
de-0.432
🎉-0.233
happy-0.231
"-0.228