INDEX

Explanations

reaction and introduction

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

 Bilg

-0.09

opi

-0.09

 EntityState

-0.09

íĹĮ

-0.09

çİ

-0.09

upe

-0.09

 toReturn

-0.09

agara

-0.08

zano

-0.08

 Harold

-0.08

POSITIVE LOGITS

 react

0.14

 reaction

0.13

 first

0.13

 React

0.12

.react

0.11

 reacts

0.11

ãģ¾ãģļ

0.11

 ë¨¼ìłĢ

0.11

React

0.11

Activations Density 0.023%