INDEX

Explanations

first instances

first presentation, time, person, program, step

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

 zuletzt

0.38

 než

0.37

 daha

0.36

 früheren

0.36

 nedenle

0.36

 mittler

0.35

 стороны

0.35

 siis

0.35

 altre

0.35

 anderen

0.34

POSITIVE LOGITS

first

0.66

 first

0.61

 पहला

0.57

FIRST

0.54

첫

0.53

 Первая

0.53

第一个

0.51

First

0.51

 primeira

0.50

 First

0.50

Activations Density 0.246%