INDEX

Explanations

ability to choose or have

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

-0.10

erton

-0.10

-0.09

 trag

-0.08

cer

-0.08

oku

-0.08

ogo

-0.08

erte

-0.08

alian

-0.08

POSITIVE LOGITS

 Ð²Ð¾Ð·Ð¼Ð¾Ð¶Ð½Ð¾ÑģÑĤÑĮ

0.15

 ability

0.14

 permission

0.13

 possibility

0.13

the

0.13

 kháº£

0.12

 moÅ¼li

0.11

 moÅ¾nost

0.11

 moÅ¼

0.10

èĥ½å¤Ł

0.09

Activations Density 0.120%