INDEX

Explanations

Jane Velez-Mitchell and JaneUnChained

myth debunked

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

 abiertas

0.49

 tiempos

0.49

 вій

0.48

 autre

0.46

 AIAA

0.46

 japonesa

0.46

藻

0.46

лянчук

0.45

 abierta

0.45

 abierto

0.45

POSITIVE LOGITS

enu

0.61

0.59

ening

0.54

0.53

0.52

iring

0.51

iled

0.51

aron

0.50

letion

0.50

LP

0.50

Activations Density 0.001%