INDEX

Explanations

sentence ending punctuation

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

 presumably

-0.11

 Rand

-0.11

ges

-0.11

eman

-0.11

 frankly

-0.11

AX

-0.11

RAND

-0.11

 Interestingly

-0.10

tb

-0.10

 surprisingly

-0.10

POSITIVE LOGITS

yon

0.09

 whenever

0.09

 wherever

0.08

 THERE

0.08

WE

0.08

strap

0.08

 THEY

0.08

 barring

0.08

zug

0.07

 chop

0.07

Activations Density 0.077%