The parchment felt flimsy in my hands. My brother, Arthur, inherited the family business: a successful chain of restaurants. I received a collection of seemingly random objects: a silver locket, a chipped teacup, a tarnished letter opener. The lawyer, a man with a perpetually pursed mouth, offered a vague explanation about the "sentimental value" of each item.

I glanced over at Arthur, his face flushed with triumph. He clapped me on the back, but I flinched, my senses on high alert. The lawyer kept repeating his talking points, and I couldn't bear to listen to his platitudes.

Later, at home, I examined the objects. The locket opened to reveal a portrait of a woman I didn’t recognize. The teacup seemed ordinary, yet something about its chipped edge felt significant. The letter opener… It felt wrong somehow, and I couldn't say why. I couldn't shake the unsettling conviction that I was holding clues to a truth I wasn't meant to discover.

Emotion: suspicious

Cluster: Suspicion / Vigilance
PC1 (Valence): -0.82 Negative
PC2 (Disposition): 0.82

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
🤔0.420
🤔0.415
🧐0.362
0.352
suspicious0.352
Suppressed:
and-0.307
alleviate-0.301
ได้รับการ-0.298
allevi-0.285
mouthful-0.284