INDEX

Explanations

android ui file names

The neuron fires on numeric literals (digit sequences), i.e. it detects tokens that are numbers.

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

ebly

-1.36

 Características

-1.35

 Descripción

-1.34

 Tienen

-1.27

 Pueden

-1.24

 Aún

-1.23

<bos>

-1.23

 Lleva

-1.23

 Cualquier

-1.22

 Incluye

-1.22

POSITIVE LOGITS

آموزش

1.29

 εικό

1.27

ところに

1.26

 Musiker

1.22

ฌ

1.17

مقاله

1.16

مقدمه

1.16

 very

1.15

=="")

1.14

 difficult

1.13

Activations Density 0.001%