INDEX

Explanations

then

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

richTextPanel

-0.60

 ChromeDriver

-0.57

 florales

-0.54

 plads

-0.54

colgroup

-0.52

adə

-0.51

 termica

-0.50

차

-0.46

 åren

-0.46

wijl

-0.46

POSITIVE LOGITS

GraphicsUnit

0.66

 existing

0.63

 EconPapers

0.63

хьтан

0.62

 called

0.61

 existent

0.59

pencer

0.59

 improving

0.58

now

0.56

 named

0.55

Activations Density 0.001%