INDEX

Explanations

requesting help or thanks

This neuron detects polite expressions of gratitude or help requests (e.g. “thanks,” “thank you,” “help me”).

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

 ProtoMessage

-1.34

 penggunaan

-1.30

 neuen

-1.29

the

-1.27

and

-1.25

ztes

-1.23

的重要

-1.23

 wichtige

-1.22

 wichtigen

-1.22

 menggunakan

-1.21

POSITIVE LOGITS

いよいよ

1.59

romeda

1.45

今年の

1.39

まずは

1.38

如果你

1.34

新型コロナ

1.30

コロナ

1.29

這個

1.29

娛樂

1.27

這是

1.27

Activations Density 0.202%