INDEX

Explanations

how

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

elemField

-0.68

EndContext

-0.65

PROTOBUF

-0.65

 незавершена

-0.58

 πως

-0.57

InjectAttribute

-0.54

enderror

-0.53

клопе

-0.53

TextAppearance

-0.52

 mergeFrom

-0.51

POSITIVE LOGITS

such

0.57

Such

0.57

'..

0.55

 legitimacy

0.55

 Such

0.53

 esternos

0.53

 AssemblyTitle

0.52

 credibility

0.49

IRQn

0.49

 such

0.48

Activations Density 0.009%