INDEX

Explanations

code here

This neuron fires on boilerplate placeholder comments that instruct the developer where to insert or add code (e.g. “Put setup code here,” “Add project-specific … rules here”).

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

-1.36

薅

-1.32

and

-1.29

-1.19

および

-1.17

-1.16

-1.14

 bota

-1.13

</h2>

-1.12

 schwä

-1.10

POSITIVE LOGITS

 skydd

1.59

iciado

1.52

cún

1.50

珦

1.47

 acá

1.40

 Üniversitesi

1.39

 alberi

1.38

 haría

1.38

小于

1.37

ưới

1.34

Activations Density 0.032%