INDEX

Explanations

import statements and code blocks

The neuron fires on documentation‐style markup tokens (e.g. reST “::” blocks, Roxygen “\{link\}” cross-refs, etc.).

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

-1.59

 його

-1.48

-1.45

菸

-1.44

for

-1.43

Even

-1.42

 utveckling

-1.41

is

-1.41

さまざまな

-1.40

了好

-1.39

POSITIVE LOGITS

 montagna

1.70

鏷

1.66

mbgg

1.64

न्ह

1.63

髕

1.63

 飾り

1.59

。『

1.59

 provar

1.58

乄

1.56

tê

1.55

Activations Density 0.001%