INDEX

Explanations

programming language keywords

The neuron fires on module‐import statements (e.g. “import” and “from” keywords, especially with “*”).

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

-1.73

 again

-1.58

 another

-1.56

its

-1.52

but

-1.50

 znovu

-1.48

쑹

-1.32

 verbind

-1.30

 incidentally

-1.28

 желание

-1.28

POSITIVE LOGITS

tanleria

1.59

STRUCTION

1.52

ühnen

1.51

)}$

1.49

)))

1.46

ضية

1.45

źdz

1.43

__);

1.43

бую

1.43

ichtete

1.42

Activations Density 0.009%