INDEX

Explanations

listing examples or purposes

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

uset

-0.10

 Ð¿Ð¾Ð±Ð°

-0.09

\tTokenName

-0.08

 chtÄĽ

-0.08

UTIL

-0.08

abled

-0.08

Ð°ÐºÐ¾Ð²

-0.08

 fork

-0.08

lr

-0.08

tablename

-0.08

POSITIVE LOGITS

ä½ľä¸º

0.17

 sebagai

0.13

çĶ¨äºİ

0.12

 aide

0.11

ahl

0.11

Used

0.11

 Used

0.11

 purposes

0.11

 ÐºÐ°ÑĩÐµÑģÑĤÐ²Ðµ

0.10

 prot

0.10

Activations Density 0.098%