INDEX

Explanations

segments of text that contain critical or emphasized information, often highlighted by variations of the word "crit" or related forms

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

 defaultstate

-0.59

MLLoader

-0.57

GIVEREF

-0.55

 Signalez

-0.54

CppMethod

-0.54

fseek

-0.51

enumi

-0.50

 jsPsych

-0.49

Nazionale

-0.49

تفصیلات

-0.49

POSITIVE LOGITS

crit

3.88

 crit

2.30

Crit

2.13

 Crit

2.02

crits

1.68

crito

1.31

crite

1.11

critto

1.11

 critters

0.94

 criter

0.81

Activations Density 0.003%