INDEX

Explanations

list markers and explanations

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

red

2.54

 vent

2.33

re

2.29

 form

2.26

 pause

2.25

 tons

2.25

 claim

2.23

 dessert

2.23

 civil

2.22

 sent

2.20

POSITIVE LOGITS

しかし

3.85

虽然

3.47

この

3.40

そして

3.34

上述

3.31

如果

3.25

その

3.20

不仅

3.16

これ

3.13

即使

3.12

Activations Density 0.045%