INDEX

Explanations

please rephrase or try again

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

 spur

-0.10

-0.09

:.|

-0.09

imp

-0.09

Ã¨res

-0.09

rint

-0.09

 indeb

-0.08

itt

-0.08

 Edison

-0.08

ÑĸÐ½Ð¾

-0.08

POSITIVE LOGITS

 repeat

0.14

Repeat

0.14

 translate

0.13

 English

0.13

 transl

0.13

 slower

0.13

 reform

0.13

 Repeat

0.12

repeat

0.12

 Ð¿Ð¾Ð²ÑĤÐ¾ÑĢ

0.12

Activations Density 0.085%