INDEX

Explanations

find followed by word or punctuation

New Auto-Interp

Configuration

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

 found

-0.13

 finding

-0.13

 findings

-0.12

found

-0.12

ember

-0.11

_found

-0.11

 fond

-0.11

Found

-0.11

izzo

-0.11

 Giles

-0.10

POSITIVE LOGITS

out

0.18

DOMNode

0.15

lay

0.14

ered

0.14

amental

0.14

 fault

0.13

 ways

0.13

 heraus

0.13

ViewById

0.13

 refuge

0.12

Activations Density 0.237%

No Known Activations