INDEX

Explanations

decide, determine, performing

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

atab

-0.10

 learned

-0.09

 Suggestions

-0.09

bet

-0.09

Emp

-0.09

iler

-0.09

 Learned

-0.09

Sr

-0.08

iller

-0.08

 disob

-0.08

POSITIVE LOGITS

 clar

0.17

 clarify

0.16

 clarification

0.15

 determining

0.14

 Clar

0.14

 decide

0.14

 clarity

0.14

 clarified

0.14

 determines

0.14

 determine

0.13

Activations Density 0.017%