INDEX

Explanations

how to use or follow

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

 input

-0.10

 Frid

-0.10

 Tall

-0.09

MSN

-0.09

è¼¸

-0.08

vul

-0.08

 Cure

-0.08

 Pall

-0.08

WithData

-0.08

(internal

-0.08

POSITIVE LOGITS

 casting

0.13

 code

0.13

ä»£çłģ

0.13

 cast

0.12

æĸ¹æ³ķ

0.12

 method

0.12

 cÃ³digo

0.11

 calling

0.10

 Calling

0.10

var

0.10

Activations Density 0.176%