INDEX

Explanations

waking and tricky situations

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

 wakes

-1.02

uxxxx

-0.98

 wake

-0.97

 ExecuteAsync

-0.93

 disambiguazione

-0.91

 woken

-0.90

 الحره

-0.89

曖昧さ回避

-0.88

 waking

-0.85

 يتيمه

-0.85

POSITIVE LOGITS

up

0.56

ままで

0.52

up

0.50

stand

0.50

ET

0.47

comp

0.46

head

0.46

us

0.45

most

0.45

Activations Density 0.267%