INDEX

Explanations

scientific studies

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

setHas

-0.54

 TextEditing

-0.54

zdo

-0.51

 qualités

-0.51

negan

-0.50

bzero

-0.49

enumi

-0.49

NESS

-0.49

 douleurs

-0.48

theoremstyle

-0.48

POSITIVE LOGITS

 Roskov

0.62

ParallelGroup

0.60

ніципа

0.59

Rüyada

0.59

fjspx

0.58

CopyWith

0.58

 FBref

0.57

<bos>

0.57

 maneuvers

0.57

ⓧ

0.56

Activations Density 0.015%