INDEX

Explanations

The

New Auto-Interp

Configuration

Prompts (Dashboard)

16,384 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

The

-1.51

-1.16

The

-1.00

↵↵

-0.90

-0.87

and

-0.84

-0.82

It

-0.81

 This

-0.78

in

-0.75

POSITIVE LOGITS

 незавершена

1.39

GEBURTSDATUM

1.22

RegressionTest

1.18

 تضيفلها

1.11

fjspx

1.09

 Мексичка

1.08

 Савезне

1.08

InjectAttribute

1.07

featureID

1.06

ształ

1.04

Activations Density 0.599%