INDEX

Explanations

concepts followed by punctuation

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

あえず

1.61

ご覧

1.51

ABLISHED

1.50

֨

1.48

 عباره

1.48

ørende

1.44

 bahut

1.43

 estiver

1.42

Nearest

1.42

 terdiri

1.41

POSITIVE LOGITS

?;

1.58

 Reddit

1.53

]);

1.49

];

1.48

));

1.42

);

1.40

?,

1.40

DIY

1.40

 Dream

1.38

 reddit

1.36

Activations Density 0.165%