INDEX

Explanations

dialogue and answering

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

 however

-1.43

にしても

-1.23

 also

-1.20

and

-1.13

 therefore

-1.10

 Therefore

-1.09

 szczeg

-1.07

argento

-1.03

 jedoch

-1.03

tanian

-1.01

POSITIVE LOGITS

Answer

1.15

当然是

1.05

<>

0.95

 answer

0.93

 answered

0.88

ローラー

0.88

Depends

0.88

 quæ

0.87

主要是

0.86

回答

0.86

Activations Density 0.013%