INDEX

Explanations

The neuron specializes in spotting keyword‐list or tag metadata (e.g. “KEYWORDS:”, “Keywords:”, hashtags, package keywords) in documents.

keywords and library names

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

為に

-1.48

of

-1.45

providedIn

-1.34

つづく

-1.33

-1.30

づけ

-1.27

にします

-1.27

ditionally

-1.25

-1.24

際には

-1.23

POSITIVE LOGITS

 două

1.66

<bos>

1.53

Willkommen

1.49

ariConfig

1.47

COMO

1.45

↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵

1.43

そして

1.42

↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵

1.41

などを

1.40

などが

1.40

Activations Density 0.022%