The coffee tasted like burnt tires. Michael swilled it, the bitter liquid doing nothing to soothe his simmering irritation. The meeting, the one about the sale, hadn't even begun, but he could already predict the platitudes, the forced smiles, the hollow promises.

He watched his colleagues, already buzzing with nervous energy, exchanging glances, murmuring in hushed tones. They seemed… eager. Or maybe just nervous. He found himself almost physically repulsed by their eagerness to pretend everything would be okay. He felt a need to break the tension, to say something cutting, but he stifled it. For now.

He adjusted his tie, feeling the unfamiliar tightness around his neck. It felt like a noose. He took a deep breath, and the air around him felt thin, the walls closing in. He'd never see these people again, and that was a welcome thought.

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