INDEX

Explanations

best, most, greatest

New Auto-Interp

Configuration

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

immer

-0.10

 gemeins

-0.10

obre

-0.10

inton

-0.09

unte

-0.09

 bekan

-0.09

ASY

-0.09

orro

-0.09

ieber

-0.09

stery

-0.09

POSITIVE LOGITS

 most

0.42

 strongest

0.39

 highest

0.38

 largest

0.37

 richest

0.36

 biggest

0.35

 finest

0.35

 brightest

0.35

 greatest

0.35

 best

0.35

Activations Density 0.523%