INDEX

Explanations

accessing object properties

The neuron fires on code identifiers in camel‐ or PascalCase (e.g. property/method names like “frame”, “layer”, “StartInfo”, “ReadAsStringAsync”, “RouteData”).

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

 that

-1.51

 such

-1.34

of

-1.33

its

-1.29

all

-1.23

一直

-1.13

于

-1.10

|};

-1.09

無し

-1.08

蛱

-1.07

POSITIVE LOGITS

了几

1.31

Nasc

1.22

 促销

1.18

動機

1.17

を

1.16

饗

1.16

かなぁ

1.14

と思いました

1.12

 "/";

1.12

möglichen

1.11

Activations Density 0.014%