INDEX

Explanations

multiple

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

keeper

-0.26

ses

-0.22

/her

-0.22

ning

-0.22

åħ¶ä¸Ńä¸Ģä¸ª

-0.21

ILA

-0.20

lands

-0.20

æŃ£å¸¸ä½¿çĶ¨

-0.20

wards

-0.20

æĢ»éĿ¢ç§¯

-0.20

POSITIVE LOGITS

 sclerosis

0.48

-choice

0.33

-step

0.29

(>

0.28

ãĢħ

0.27

/m

0.26

 ways

0.26

-layer

0.25

-thirds

0.25

-headed

0.25

Activations Density 0.064%