INDEX

Explanations

code strings and context

This neuron is triggered by string literals containing path‐style slashes (e.g. file system or URL/API paths).

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

-1.50

帰る

-1.35

-1.32

’

-1.30

not

-1.28

 أيض

-1.26

什么呢

-1.26

罍

-1.23

en

-1.23

こっち

-1.22

POSITIVE LOGITS

Ꮑ

1.55

apatkan

1.48

 Wirken

1.38

But

1.34

 этого

1.30

Questa

1.29

 estekak

1.27

Život

1.27

ucción

1.26

raulic

1.24

Activations Density 0.114%