INDEX

Explanations

references to various highly specific robot or weapon arrangements and configurations

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

enderror

-0.63

moso

-0.56

 Franz

-0.56

uze

-0.54

MessageTagHelper

-0.53

__':

-0.52

Franz

-0.52

ASN

-0.52

ghis

-0.51

finalize

-0.50

POSITIVE LOGITS

Ry

3.27

Ry

2.84

ry

1.80

RY

1.80

 Ryder

1.12

Ryu

0.96

Ryd

0.79

 Ры

0.78

Ryu

0.73

Ryo

0.73

Activations Density 0.002%