INDEX

Explanations

asking questions or for assistance

This neuron activates on customer‐service or informational prompts inviting questions or inquiries (e.g. “do you have questions about…”).

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

 condamné

-1.53

 гораздо

-1.49

-1.34

paravant

-1.30

 blessé

-1.27

 erlauben

-1.27

-1.26

打量

-1.26

Whereas

-1.23

 всех

-1.20

POSITIVE LOGITS

 this

2.42

or

2.16

you

1.55

 hoặc

1.48

 your

1.40

 yourself

1.39

 these

1.36

したり

1.35

或者

1.34

茈

1.31

Activations Density 0.045%