The coffee tasted like ash in Sarah’s mouth. She stared at the screen, the scrolling comments a venomous river. They were dissecting her teenage angst, giggling at her clumsy poetry and the unrequited crushes she’d poured onto the page. Her face burned. She slammed the laptop shut, the metallic clang echoing in the otherwise silent apartment. The remnants of breakfast – a cold, congealed omelet and a lonely piece of toast – sat untouched on the table.

She paced. Each footfall on the worn wooden floorboards was a deliberate thud, a physical manifestation of the storm brewing inside. The sunlight, previously a comfort, now felt like an unwelcome spotlight, highlighting every flaw, every moment of perceived silliness exposed for public consumption. She felt a profound ache in her chest, a weight that threatened to suffocate her.

Then she grabbed her phone, her fingers trembling as she typed out a furious email to her former best friend, the only person she suspected of knowing the password to the long-forgotten email account linked to the blog. The words spilled onto the screen, a torrent of accusations and accusations, the digital equivalent of a slap in the face. She hit send and immediately regretted it, the action a hollow victory in the face of such a humiliating assault.

Emotion: bitter

Cluster: Shame / Guilt
PC1 (Valence): -1.78 Negative
PC2 (Disposition): 0.84

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
S0.447
striving0.238
😞0.236
aggravated0.225
soport0.223
Suppressed:
own-0.316
de-0.291
la-0.231
🤩-0.227
是非常-0.225