INDEX

Explanations

javascript code imports

This neuron fires on short, single-character identifiers (like “r”, “l”, “a”, “s”) typical of minified JavaScript variable names.

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

 when

-1.35

 where

-1.23

 only

-1.19

for

-1.17

 dato

-1.13

Probably

-1.09

When

-1.08

Karakteristik

-1.05

Have

-1.02

 just

-1.02

POSITIVE LOGITS

鹩

1.23

vement

1.19

sonian

1.18

{}{

1.14

)$,

1.13

WEBPACK

1.13

 UNIVERS

1.12

ucine

1.09

THON

1.09

पि

1.08

Activations Density 0.001%