INDEX

Explanations

code delimiters and structure

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 preparations

1.18

 thinkers

1.18

 enabling

1.17

↵↵↵↵↵↵↵↵

1.17

 alone

1.15

 forced

1.15

 freely

1.15

 freed

1.15

 transformed

1.15

 greatly

1.14

POSITIVE LOGITS

 tör

1.47

<0xF3>

1.41

 humedad

1.36

 vele

1.34

 வழக்கு

1.34

 mostrado

1.32

 mindig

1.32

 laag

1.30

 vég

1.29

 vatten

1.28

Activations Density 0.003%