INDEX

Explanations

increasing progressively

Explanation of neuron 4 behavior: the main thing this neuron does is find content-bearing nouns and verbs—especially technical or informational terms—in otherwise function-word–heavy text.

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

 heridas

0.71

 Toxicology

0.68

 Tumor

0.66

 eterna

0.64

 tajem

0.64

 CHEMISTRY

0.64

 Archive

0.63

 luciferase

0.63

 Preserve

0.62

 сохранения

0.62

POSITIVE LOGITS

 dần

1.47

 gradually

1.40

 постепенно

1.40

 progressivement

1.37

逐漸

1.34

逐渐

1.32

逐步

1.26

 gradual

1.25

 постепен

1.18

渐渐

1.15

Activations Density 1.390%