INDEX

Explanations

proportional or inverse relationship

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

boo

-0.11

 ASAP

-0.10

lover

-0.09

 hoped

-0.09

sod

-0.09

"struct

-0.09

tik

-0.09

 preferred

-0.08

 tact

-0.08

 cooper

-0.08

POSITIVE LOGITS

 proportional

0.35

 invers

0.34

 proportion

0.25

 inverse

0.23

 correlate

0.22

 correlated

0.22

 correlates

0.21

 dependent

0.18

Inverse

0.18

 correl

0.18

Activations Density 0.465%