INDEX

Explanations

gh followed by specific endings/names

New Auto-Interp

Configuration

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

ucer

-0.11

/docs

-0.10

chied

-0.10

oppins

-0.10

opoulos

-0.10

aking

-0.10

sheets

-0.10

oles

-0.10

ICLES

-0.09

asel

-0.09

POSITIVE LOGITS

rah

0.10

osh

0.09

irma

0.09

apos

0.09

idor

0.09

strup

0.08

Gh

0.08

zion

0.08

yll

0.08

urar

0.08

Activations Density 0.008%

No Known Activations