INDEX

Explanations

possessives ending in 's

New Auto-Interp

Configuration

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Toby

-0.11

ase

-0.11

misc

-0.10

 past

-0.10

 Zack

-0.10

 Hoffman

-0.10

ï¸ı

-0.10

ups

-0.10

illus

-0.10

 utility

-0.09

POSITIVE LOGITS

 recuper

0.09

 Harris

0.09

ned

0.08

 descend

0.08

 dire

0.08

achen

0.08

bine

0.08

 Solo

0.07

elim

0.07

 Barcl

0.07

Activations Density 0.001%