INDEX

Explanations

may be superseded

The neuron detects uppercase boilerplate headers or notice text (e.g. all-caps transcript or legal disclaimer lines).

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

 their

-1.76

 Using

-1.35

 using

-1.31

 blatantly

-1.30

 with

-1.29

in

-1.28

 really

-1.26

 especially

-1.26

 Having

-1.24

 Throughout

-1.23

POSITIVE LOGITS

titial

1.47

either

1.46

generally

1.39

 alcuni

1.37

believe

1.34

usually

1.32

therefore

1.31

igth

1.31

often

1.30

typically

1.30

Activations Density 0.209%