INDEX

Explanations

non-english text

different types and categories

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

0.53

Loc

0.47

blast

0.45

teacher

0.45

{}\

0.44

{~

0.44

unit

0.43

tower

0.43

MUST

0.43

ninja

0.42

POSITIVE LOGITS

länd

0.54

ländischen

0.53

 څرنګوالی

0.50

ernsey

0.48

 краї

0.48

朐

0.48

 Росії

0.46

 INSEE

0.45

ه‌ای

0.45

 সালে

0.45

Activations Density 0.001%