INDEX

Explanations

code structures and commands

the literal number zero in code or mathematical expressions.

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

 gier

-1.55

 tranſ

-1.47

 spada

-1.45

 ondu

-1.44

薅

-1.43

probably

-1.41

 konkret

-1.40

 jaka

-1.38

 atak

-1.38

peut

-1.38

POSITIVE LOGITS

as

1.88

and

1.78

 with

1.66

 made

1.60

 With

1.39

by

1.38

 helped

1.30

 after

1.28

1.27

 vital

1.27

Activations Density 0.012%