INDEX

Explanations

produced by

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Å¼y

-0.07

 Ð·Ð°ÑĤÑı

-0.07

ounty

-0.07

kopf

-0.07

æ¯Ķè¾ĥéķ¿

-0.07

à¹ĥà¸Ļà¸ķà¸±à¸§

-0.07

æĺ¯åĲ¦éľĢè¦ģ

-0.07

dw

-0.07

å¦ĤæŀľéľĢè¦ģ

-0.07

ectes

-0.07

POSITIVE LOGITS

 ranges

0.12

 range

0.11

çļĦèĮĥåĽ´

0.11

èĮĥåĽ´

0.11

èĮĥåĽ´ä¸º

0.10

 Range

0.10

range

0.10

_range

0.10

Range

0.09

ranges

0.09

Activations Density 0.012%