INDEX

Explanations

number of

part of words in many languages

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

 trailers

0.42

 fashioned

0.42

 kuin

0.42

 priced

0.41

 isotopes

0.41

nq

0.40

 gasoline

0.40

 kenn

0.40

 nourished

0.40

 scores

0.40

POSITIVE LOGITS

の一部

0.52

 அதிகரி

0.48

}(

0.47

 graduellement

0.45

 మరింత

0.45

ของ

0.45

 kvůli

0.44

 zejména

0.44

 של

0.43

ექს

0.43

Activations Density 0.001%