INDEX

Explanations

opening parenthesis, bracket, brace

language model

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

assaytype

0.47



0.46

🏪

0.46

 Caball

0.46

 desktops

0.45

 alfalfa

0.44

🕘

0.43

 invit

0.42

zeptember

0.42

⚉

0.42

POSITIVE LOGITS

xrightarrow

0.79

ensuremath

0.75

 bracket

0.74

bracket

0.72

rrbracket

0.72

mathb

0.72

 brackets

0.70

Bracket

0.70

 Bracket

0.69

Declare

0.68

Activations Density 0.004%