INDEX

Explanations

conjunctions and conditional words

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

 якого

1.27

和我

1.20

和他

1.16

를

1.13

および

1.11

 మారి

1.10

álního

1.09

を発表

1.08

 تواند

1.08

oprotein

1.07

POSITIVE LOGITS

 there

2.73

 although

2.40

 despite

2.30

 unless

2.15

there

2.12

 whenever

2.07

if

2.05

 when

1.97

 unlike

1.96

 while

1.95

Activations Density 5.519%