INDEX

Explanations

citations and references

The neuron fires on numeric tokens (strings of digits such as years, reference numbers, and page counts).

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

-1.52

 ͜ʖ

-1.28

),

-1.22

作詞

-1.16

useNewUrlParser

-1.16

 fernández

-1.10

 garcía

-1.09

ㄟ

-1.09

jerg

-1.09

zkę

-1.08

POSITIVE LOGITS

1.29

to

1.23

↵

1.16

sce

1.04

 czasem

1.02

1.00

jual

0.96

 spannende

0.94

 originalen

0.93

°;

0.93

Activations Density 0.005%