INDEX

Explanations

historical date ranges or years

The neuron fires on numeric tokens that form years or year‐range expressions (e.g. the digits and hyphens in “1980–1987”).

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

one

-1.50

new

-1.36

two

-1.33

so

-1.32

now

-1.30

 this

-1.29

 different

-1.28

 like

-1.27

by

-1.23

 previous

-1.23

POSITIVE LOGITS

 リュック

1.22

 دقيقة

1.14

 スイッチ

1.13

 Quando

1.06

 لهذه

1.05

 เปิด

1.05

 entweder

1.04

 图解

1.04

 ภาพ

1.02

Interes

1.02

Activations Density 0.043%