INDEX

Explanations

found the next word

The neuron flags tokens written in Cyrillic (i.e. Russian-language text).

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

־ה

-1.42

 three

-1.28

ÈNE

-1.27

 five

-1.19

 encara

-1.19

 היתה

-1.18

 triliun

-1.17

 caucho

-1.17

==""){

-1.16

founder

-1.16

POSITIVE LOGITS

我看

1.28

यह

1.21

 данный

1.15

 Gehir

1.14

暖かい

1.13

 possible

1.10

想必

1.10

 mógł

1.08

 leech

1.07

 việt

1.05

Activations Density 0.019%