INDEX

Explanations

so

New Auto-Interp

Configuration

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

uran

-0.27

ESA

-0.27

 stem

-0.27

arranty

-0.26

ebra

-0.26

ãģªãģľ

-0.25

ODO

-0.25

odes

-0.25

ÑģÑĤÐ°Ð½Ð¾Ð²

-0.24

ncy

-0.24

POSITIVE LOGITS

 Actions

0.28

åĲ¬äºĨ

0.27

æ²¾

0.26

aked

0.25

æ´ĩ

0.25

ħ§

0.24

æ¹¿

0.24

Ð¾ÑģÑĮ

0.24

 Fortress

0.24

%[

0.24

Activations Density 0.002%