INDEX

Explanations

all except one

The neuron activates on occurrences of the word “except,” marking exclusion clauses.

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

and

-1.79

by

-1.63

 though

-1.60

or

-1.57

in

-1.52

 This

-1.48

 then

-1.46

owohl

-1.44

 jeweils

-1.39

xtext

-1.36

POSITIVE LOGITS

the

2.11

 هستند

1.92

を開始

1.69

 グリル

1.63

﹍﹍

1.62

 pérd

1.57

 huel

1.56

1.52

 télécommande

1.51

醃

1.51

Activations Density 0.014%