INDEX

Explanations

private data and internal identifiers

The neuron fires strongly on empty‐string tokens that mark paragraph or section breaks (i.e. blank lines/newline separators).

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

0.80

 hInstance

0.79

 bassist

0.78

֜

0.73

 Legend

0.72

 Routledge

0.72

 einzelnen

0.71

 সত্যিকারের

0.71

 جلو

0.71

ie

0.70

POSITIVE LOGITS

ك

0.87



0.77

物の

0.76

暏

0.76

的我

0.74

debris

0.72

ruby

0.71

cherry

0.71

ताच

0.71

 स्ट्रीट

0.71

Activations Density 0.000%