INDEX

Explanations

qualities and attributes

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

_Lean

-0.11

 láº«n

-0.11

_Tis

-0.11

ëĿ¼ëıĦ

-0.09

çĶ

-0.09

ãĥ¬ãĤ¹

-0.09

gebra

-0.09

AndGet

-0.08

_EMIT

-0.08

ouz

-0.08

POSITIVE LOGITS

or

0.18

ä¸Ķ

0.15

and

0.15

 rather

0.14

/icon

0.11

ãĢģ

0.10

 stuff

0.10

 amounts

0.10

0.09

Activations Density 0.177%