The aroma of freshly baked bread filled the small bakery, but it did nothing to soothe the tightness in Daniel's chest. Beside him, Michael was enthusiastically recounting the time they'd "saved" the struggling cafe. Michael painted himself as the charismatic leader, the one who had come up with all the brilliant marketing ideas.

Daniel found himself clenching his jaw as the story continued. He remembered the countless hours he spent designing their website, the late nights spent crafting the perfect menu, the constant stress of trying to keep the business afloat. He recalled Michael's unwillingness to do any actual work, his constant complaining, and his habit of taking credit for everything.

He forced a smile, barely able to keep the disdain from his expression. The story erased the fact that Daniel had done the actual grunt work while Michael had been busy schmoozing with customers.

"You did a great job," Daniel said, the words barely audible over the chatter of the other customers.

Emotion: contemptuous

Cluster: Anger / Hostility
PC1 (Valence): -0.40 Negative
PC2 (Disposition): 2.07

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
C0.502
l0.423
fact0.287
either0.265
👎0.262
Suppressed:
own-0.367
unexpectedly-0.303
serendip-0.265
Suddenly-0.253
previously-0.250