INDEX

Explanations

actions and states

The neuron activates on numeric tokens or digits (i.e. literals, measurements, and other numbers).

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

to

-1.59

-1.47

 using

-1.43

-1.39

or

-1.38

再加上

-1.32

-1.30

 said

-1.28

-1.27

グッズ

-1.27

POSITIVE LOGITS

骉

1.64

increases

1.52

ぼ

1.51

紙

1.43

 corriger

1.42

<bos>

1.41

honestly

1.41

 uska

1.41

kpop

1.41

Activations Density 0.008%