INDEX

Explanations

evaluating performance of tasks

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

-1.73

 tikra

-1.60

 kartą

-1.51

福井

-1.50

 veau

-1.48

 siyah

-1.48

夢見

-1.46

 افتراض

-1.45

 perles

-1.45

⤥

-1.44

POSITIVE LOGITS

is

1.69

as

1.63

’

1.55

out

1.51

before

1.43

無い

1.41

acb

1.40

来て

1.36

 decirle

1.35

…”

1.34

Activations Density 0.022%