INDEX

Explanations

systems

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

Structure

-0.68

 structures

-0.67

structure

-0.65

 structure

-0.63

 restructure

-0.60

 restructuring

-0.60

tagext

-0.59

TargetException

-0.57

 GenerationType

-0.55

ndrome

-0.55

POSITIVE LOGITS

to

0.61

 ModelExpression

0.61

ⓧ

0.54

SizeMode

0.52

ambung

0.52

IsContent

0.51

 informed

0.50

at

0.50

dis

0.49

 שוליים

0.49

Activations Density 0.024%