INDEX

Explanations

Research papers

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

مصادر

-0.70

FontOfSize

-0.66

ally

-0.58

‬‬

-0.58

**********

-0.54

‭

-0.54

nal

-0.52

 portion

-0.52

mstyle

-0.51

abetes

-0.51

POSITIVE LOGITS

:✨

0.77

niająca

0.60

findpost

0.59

InitVars

0.56

0.53

dymyr

0.52

 dAtA

0.52

expandindo

0.52

 дописавши

0.50

κτή

0.50

Activations Density 0.030%