INDEX

Explanations

one-stop-shop, complete, single

New Auto-Interp

Top Features by Cosine Similarity

Configuration

Prompts (Dashboard)

10,000 prompts, 128 tokens each

Dataset (Dashboard)

lmsys/lmsys-chat-1m

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

çıł

-0.09

æģ

-0.08

ways

-0.08

pickle

-0.08

âĢŀP

-0.08

igner

-0.08

words

-0.08

iá»ģn

-0.08

 Sommer

-0.08

sch

-0.07

POSITIVE LOGITS

one

0.38

ONE

0.31

one

0.30

One

0.29

One

0.28

 satu

0.25

_one

0.25

ONE

0.23

-one

0.23

.one

0.22

Activations Density 0.094%