INDEX

Explanations

directory structures and code formatting

The neuron spotlights ASCII‐art “tree” diagram lines (the box‐drawing branch characters like ├──, └──, slashes and backslashes used to depict directory or diagram structure).

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

 THOSE

-1.40

ができる

-1.38

){$\

-1.27

кож

-1.25

 anderen

-1.24

Altre

-1.24

跛

-1.23

༘

-1.23

脓

-1.23

LEGGI

-1.20

POSITIVE LOGITS

 three

1.84

two

1.77

six

1.70

one

1.55

 four

1.53

 seven

1.48

for

1.36

zied

1.34

 first

1.32

 five

1.31

Activations Density 0.043%