INDEX

Explanations

wide range of

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

äºĽ

-0.10

IEWS

-0.08

igner

-0.08

 plentiful

-0.08

VERTISEMENT

-0.08

 å·Ŀ

-0.08

anean

-0.08

ìĽħ

-0.08

iba

-0.08

bracht

-0.08

POSITIVE LOGITS

 range

0.19

 wide

0.17

 eine

0.16

 variety

0.15

 ÑĪÐ¸ÑĢÐ¾Ðº

0.14

 ranges

0.13

een

0.12

range

0.12

 widest

0.12

Activations Density 0.205%