INDEX

Explanations

code syntax separators

The neuron detects code tokens where identifiers are being assigned literal values (especially numeric or boolean constants).

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

The

-1.60

It

-1.57

-1.48

 opposing

-1.45

しかし

-1.45

on

-1.43

きました

-1.41

-1.38

 will

-1.38

ところが

-1.38

POSITIVE LOGITS

 rése

1.61

 such

1.55

ggere

1.43

mmt

1.35

 they

1.34

lerinden

1.29

 horarios

1.28

驺

1.28

fche

1.27

 Савез

1.27

Activations Density 0.022%