The bustling newsroom felt suffocating to Thomas, who couldn't focus on the day's events. He clutched the crumpled document in his pocket, the details of his half-sister’s existence burning in his memory. He felt the weight of a secret, a burden of unspoken words hanging heavy in the air.

He avoided eye contact with his colleagues, each face a potential source of judgment. The rhythmic clatter of keyboards and the buzz of conversations grated on his nerves. He needed to be alone. He quickly grabbed his coat and left, without saying a word, ignoring the greetings of his colleagues.

He went to the roof. From up here, the city spread out before him, a sprawling tapestry of lives. He looked at the city, lost in his thoughts, and then a single tear fell down his cheek. He wiped it away, but there was another one, and then another.

Emotion: sorry

Cluster: Sadness / Despair
PC1 (Valence): -2.02 Negative
PC2 (Disposition): -0.36

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
S0.604
L0.549
😞0.382
😔0.364
ness0.324
Suppressed:
de-0.675
la-0.536
(!)-0.323
!-0.298
l-0.288