INDEX

Explanations

suggesting options or providing help

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

 sobie

-0.09

 Curry

-0.09

Lau

-0.08

Await

-0.08

 Govern

-0.08

 Specify

-0.08

_interaction

-0.08

 Ð¿Ð¾Ð»ÑĮÐ·

-0.08

cmds

-0.08

Own

-0.08

POSITIVE LOGITS

 provide

0.26

æıĲä¾Ľ

0.26

 help

0.24

 providing

0.23

 offer

0.23

 provides

0.22

å¸®

0.21

provide

0.21

 cung

0.21

 answer

0.21

Activations Density 0.242%