INDEX

Explanations

for automobiles or elections

This neuron detects the volume number in legal case citations (the leading numeric reporter volume).

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

szt

-1.39

 japones

-1.38

átiles

-1.29

in

-1.26

 voks

-1.23

yym

-1.22

<bos>

-1.20

ıyorum

-1.16

 manualidades

-1.14

بور

-1.13

POSITIVE LOGITS

1.39

or

1.33

1.31

1.28

 REALLY

1.27

ches

1.27

1.21

‐

1.21

 этими

1.19

1.18

Activations Density 0.000%