INDEX

Explanations

type annotations in code

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

があり

-1.46

 when

-1.34

</h2>

-1.33

all

-1.27

 that

-1.25

 actually

-1.20

 where

-1.19

參考文獻

-1.15

 other

-1.15

did

-1.13

POSITIVE LOGITS

ごめん

1.72

久し

1.67

あー

1.51

ほんと

1.48

ほら

1.43

マジで

1.41

感じが

1.38

すっ

1.37

さぁ

1.36

ホント

1.34

Activations Density 0.014%