INDEX

Explanations

and provides, have had

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

åĢ

-0.12

immers

-0.11

-0.10

INST

-0.10

ddit

-0.10

//:

-0.10

ustr

-0.10

programming

-0.09

thesis

-0.09

LOBAL

-0.09

POSITIVE LOGITS

 applications

0.19

 Applications

0.17

 applicable

0.15

applications

0.15

Applications

0.14

 applic

0.13

åºĶçĶ¨

0.13

 application

0.13

 practical

0.12

 applied

0.12

Activations Density 0.059%