INDEX

Explanations

error message

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

to

-0.94

ceedings

-0.93

cluir

-0.92

yes

-0.91

做出

-0.87

 melalui

-0.87

Yes

-0.86

inchen

-0.85

gmail

-0.84

 available

-0.84

POSITIVE LOGITS

 message

1.84

 messages

1.46

 when

1.35

messages

1.25

 stating

1.16

message

1.10

saying

1.09

msgs

1.06

 thrown

1.06

 Message

1.06

Activations Density 0.186%