INDEX

Explanations

length and substring operations

The neuron detects calls to string‐manipulation methods (e.g. find, substring, split, indexOf, length).

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

 they

-1.71

up

-1.36

还有

-1.33

ooooo

-1.23

 there

-1.23

還有

-1.20

 Their

-1.16

 rade

-1.16

之力

-1.16

 maxime

-1.16

POSITIVE LOGITS

IsOptional

1.41

 after

1.23

 svého

1.18

 bekämp

1.16

*}[

1.15

 navnet

1.13

or

1.10

actrice

1.10

 nervously

1.09

 trennen

1.09

Activations Density 0.047%