INDEX

Explanations

following specific words

switch header how reflect simple architecture computational

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

ر

0.59

ي

0.56

崱

0.56

ال

0.54

瞍

0.49

 feminists

0.47

 kleur

0.46

रु

0.45

猕

0.45

 наше

0.44

POSITIVE LOGITS

 ಪ್ರದೇಶ

0.46

hao

0.43

 bypass

0.43

toa

0.42

nullable

0.42

ameni

0.41

eli

0.40

effectuer

0.40

fd

0.40

heet

0.40

Activations Density 0.000%