INDEX

Explanations

introducing numbered sections

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

 fevereiro

1.41

৭

1.36

ồi

1.35

૧

1.34

テン

1.34

২

1.33

iezan

1.29



1.29

 chiếc

1.28

 coisas

1.28

POSITIVE LOGITS

 point

0.98

 implies

0.95

 requiring

0.91

 related

0.90

 suggestion

0.90

 points

0.90

 induced

0.88

 ultimately

0.86

 proceed

0.85

 proceeds

0.84

Activations Density 0.101%