INDEX

Explanations

code paths and errors

The neuron is looking for runs of non-letter delimiter characters—e.g. sequences of asterisks or repeated punctuation like "::"—that mark separators or dividers in code or data.

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

</h6>

-1.80

ي

-1.74

</h3>

-1.73

４

-1.64

It

-1.58

２

-1.57

You

-1.55

That

-1.54

/**

-1.53

ть

-1.50

POSITIVE LOGITS

the

1.53

 größe

1.45

كونات

1.41

 rechte

1.41

除此

1.39

✑

1.37

 schim

1.37

 débuts

1.35

needs

1.35

Activations Density 0.004%