The air in the family home felt heavy, thick with the silence of the recently departed. John dragged the box across the floor, its contents rattling ominously. He should have been anywhere else. He had a million things he could be doing. He hated this.

He found the letter tucked within a Bible. His mother’s resignation. He knew it was coming. He already knew the story. He opened the letter, his face set. He read the words. The words he had heard a hundred times.

The carefully chosen phrases. The veiled criticisms. He felt a dull ache in his chest. It felt fake, like she was trying too hard. He didn’t want to read it. He knew the whole story. He wanted to go. He just needed to be done with this. With a sigh, he tossed the letter aside. It was time for a break.

Emotion: exasperated

Cluster: Anger / Hostility
PC1 (Valence): -0.99 Negative
PC2 (Disposition): 1.23

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
S0.435
l0.378
P0.331
que0.320
la0.269
Suppressed:
own-0.433
gradually-0.254
soon-0.248
unfolding-0.216
joyful-0.216