INDEX

Explanations

designed implementation intended policy effect

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

 hän

-1.36

Typically

-1.20

 jejich

-1.17

*///

-1.10

bestimm

-1.10

Válasz

-1.09

 بازیگر

-1.08

ׇ

-1.08

dä

-1.08

変わらず

-1.07

POSITIVE LOGITS

 designed

1.59

ּוֹ

1.57

 implementation

1.54

 intended

1.47

it

1.40

 implemented

1.34

 itself

1.31

 policy

1.29

its

1.27

 effect

1.24

Activations Density 0.076%