INDEX

Explanations

help, forget

New Auto-Interp

Configuration

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

to

-0.57

كويكب

-0.57

 help

-0.48

canestro

-0.44

 Bourgoin

-0.41

тельство

-0.40

niosek

-0.38

 passphrase

-0.36

有哪些

-0.36

ambung

-0.35

POSITIVE LOGITS

AccessorTable

0.84

')['

0.67

tists

0.62

 Pioneers

0.61

msgTypes

0.61

andExpect

0.59

takers

0.58

ConstraintMaker

0.58

enumi

0.58

fail

0.58

Activations Density 0.010%