INDEX

Explanations

code delimiters and data types

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 reducing

-0.93

 those

-0.91

as

-0.86

 including

-0.85

 improving

-0.85

 name

-0.82

How

-0.82

减

-0.82

 what

-0.81

 without

-0.81

POSITIVE LOGITS

 Integer

1.36

intval

1.31

Integer

1.31

 integers

1.30

 integer

1.27

 parseInt

1.27

 strconv

1.20

parseInt

1.16

 intval

1.15

integer

1.13

Activations Density 0.019%