INDEX

Explanations

representedtestholdermarkerofftrip

user prompts

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

<0x0D>

0.49

 Maiden

0.46

রাতে

0.45

0.44

 America

0.43

 inaugural

0.43

 maiden

0.43

 Displacement

0.42

 subspaces

0.42

 Nights

0.41

POSITIVE LOGITS

кар

0.55

및

0.52

ជំងឺ

0.51

че

0.51

ងឺ

0.51

疳

0.50

ﾛ

0.49

دة

0.49

すべての

0.48

ロウ

0.48

Activations Density 0.001%