INDEX

Explanations

method calls or section headers

The neuron fires on code tokens invoking retrieval methods—particularly calls to “get(…)” or “find(…)” (including “findAll(…)”).

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

 parachoque

-1.42

giày

-1.37

 estadounid

-1.30

 GRATU

-1.27

tront

-1.26

 مناط

-1.24

 hamil

-1.24

 jati

-1.24

 tulum

-1.24

 sentier

-1.23

POSITIVE LOGITS

or

1.28

There

1.24

1.23

It

1.19

1.18

 There

1.17

Now

1.16

1.15

Activations Density 0.001%