INDEX

Explanations

code-related keywords

This neuron activates on tokens that contain numeric characters (i.e. identifiers or words with digits in them).

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

 that

-1.72

its

-1.62

any

-1.48

 what

-1.38

 their

-1.36

 During

-1.33

 from

-1.32

in

-1.30

 when

-1.28

 Both

-1.24

POSITIVE LOGITS

は

2.16

を

1.88

が

1.65

や

1.49

PASO

1.43

では

1.41

これは

1.34

から

1.34

 monstru

1.28

 Collège

1.28

Activations Density 0.002%