INDEX

Explanations

paths in code

The neuron activates on strings of digits—i.e. numeric tokens.

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

 माउंट

0.59

WOR

0.59

↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵

0.57

 Clean

0.57

 Dojo

0.57

猕

0.57

লাইন

0.57

 Better

0.56

 модуль

0.56

 অভিব

0.55

POSITIVE LOGITS

 பொழுது

0.55

moires

0.53

বেগ

0.53

Extensions

0.52

 equidistant

0.51

 নিকট

0.50

확

0.49

աց

0.49

 தொலைக்கா

0.49

 فشار

0.49

Activations Density 0.000%