INDEX

Explanations

numbers, dates, and identifiers

The neuron strongly activates on numeric tokens (table or code values), i.e. it detects 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

感じで

-1.25

tième

-1.20

事で

-1.20

為に

-1.19

 remor

-1.18

•.

-1.17

instead

-1.15

ülő

-1.14

Ｙ

-1.13

 combinatie

-1.12

POSITIVE LOGITS

not

1.71

in

1.70

astify

1.38

no

1.37

がある

1.32

now

1.30

などは

1.30

to

1.29

があります

1.29

at

1.22

Activations Density 0.028%