INDEX

Explanations

the word 'nor' with high activation, and in some cases also the word 'or.'

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

出版年

-0.90

or

-0.77

<bos>

-0.76

或

-0.69

 &___

-0.68

 или

-0.66

ή

-0.60

سطس

-0.59

或

-0.59

OR

-0.59

POSITIVE LOGITS

CloseOperation

0.60

thums

0.56

acebook

0.53

Groet

0.53

 platelets

0.52

plash

0.52

transpiration

0.52

chiali

0.52

bertas

0.50

diferentes

0.50

Activations Density 0.498%