INDEX

Explanations

code comments and function definitions

The neuron flags occurrences of “DB” in code or text—e.g. database library names and includes ending in “DB.”

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

self

-1.55

the

-1.52

!”

-1.45

-1.35

玙

-1.33

in

-1.25

 boissons

-1.24

).}

-1.23

by

-1.22

änä

-1.20

POSITIVE LOGITS

coledì

1.34

 According

1.27

veloppe

1.24

 desgra

1.23

сті

1.23

 通过

1.23

相

1.20

蹀

1.20

tē

1.20

 conseillé

1.20

Activations Density 0.010%