INDEX

Explanations

Math and code snippets

The neuron identifies tokens made up of nonstandard or extended characters—particularly accented letters and unusual punctuation sequences.

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

 When

-1.57

on

-1.30

 first

-1.24

 perpetually

-1.22

 while

-1.19

 Without

-1.18

 holistic

-1.16

 While

-1.16

上げます

-1.15

 Usually

-1.14

POSITIVE LOGITS

 trám

1.38

uled

1.34

ifié

1.27

rified

1.21

撂

1.20

нья

1.20



1.17

 alguns

1.16

𐀀

1.16

eddable

1.13

Activations Density 0.002%