The dinner party was underway, and Emily was struggling to keep it together. She'd worked hard to create a welcoming atmosphere, but the secret of the house was a heavy weight. She'd tried to bury the feeling of it by spending hours arranging the candles, choosing the perfect playlist. The guests were all friends, people she wanted to impress.

She poured the wine, her hand shaking slightly. Someone mentioned the local history, and a woman asked, "Didn't someone…die here?" Emily's stomach dropped. She forced a laugh. "Oh, that old wives' tale?" She tried to sound breezy, but her voice cracked.

She found herself talking non-stop, eager to fill the silence, to drown out the unspoken question in the air. She made sure that everyone had a full glass, and took extra pains to laugh at every joke. She felt a burning pressure on her cheeks, each syllable felt strained, as if she were carrying a secret that was eating her away from the inside.

Emotion: self-conscious

Cluster: Embarrassment
PC1 (Valence): -1.03 Negative
PC2 (Disposition): 0.52

Role in Research

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

Logit Lens (Layer 40)

Tokens promoted/suppressed when the self-conscious vector is projected through the unembedding matrix.

Promoted:
S0.307
C0.298
ness0.273
尴尬0.263
😶0.242
Suppressed:
que-0.399
de-0.266
triumphant-0.266
আনন্দের-0.247
ايا-0.238