INDEX

Explanations

attempting to take action

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

and

-2.88

縞

-2.30

และ

-2.06

鶫

-2.06

 creat

-2.06

(!__

-2.03

出了一

-1.98

ㆆ

-1.95

不觉

-1.95

鸝

-1.94

POSITIVE LOGITS

by

2.45

 There

2.42

 While

2.39

2.28

 devise

1.98

 biggest

1.92

 Using

1.91

</em>

1.83

～～～～

1.83

but

1.82

Activations Density 0.025%