INDEX

Explanations

negative conditions or states

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

 both

-1.52

２

-1.41

４

-1.41

 beiden

-1.39

 både

-1.38

条件

-1.36

許多

-1.36

 doivent

-1.34

许多

-1.32

１

-1.32

POSITIVE LOGITS

 adequate

1.76

 appropriate

1.70

 properly

1.66

 comfortable

1.62

its

1.60

 where

1.59

 understand

1.57

 correct

1.57

怎么回事

1.55

 correctly

1.53

Activations Density 0.146%

No Known Activations