INDEX

Explanations

attempts and efforts

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

venge

-0.10

usan

-0.09

 Laud

-0.09

 resid

-0.08

 circumstance

-0.08

enate

-0.08

lug

-0.08

ugged

-0.08

_CPP

-0.08

olly

-0.08

POSITIVE LOGITS

 attempts

0.65

 efforts

0.51

 Attempts

0.49

 attempt

0.48

 attempted

0.42

Attempts

0.41

 Attempt

0.36

Attempt

0.35

 tries

0.34

attempt

0.34

Activations Density 0.531%