INDEX

Explanations

opening files and code elements

This neuron activates on literal filename or database‐connection string arguments in code (e.g. the path passed to open(), connect(), etc.).

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

เป็น

-1.10

is

-1.04

nees

-1.02

前は

-1.02

ru

-0.98

这种

-0.98

がありますが

-0.96

นี้

-0.96

ens

-0.96

rie

-0.96

POSITIVE LOGITS

 тук

1.30

 санкт

1.27

 всички

1.27

ܽ

1.23

えっ

1.23

 nastro

1.21

zbęd

1.20

 москве

1.20

 obuv

1.20

pecabe

1.20

Activations Density 0.022%