INDEX

Explanations

lack of care or regard

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

ëª»

-0.10

-0.09

ENTE

-0.09

ä¸įå¾Ĺ

-0.08

exc

-0.08

(always

-0.08

uÄį

-0.08

UFFER

-0.08

Parms

-0.08

POSITIVE LOGITS

 disregard

0.25

 care

0.24

 indifference

0.22

 disreg

0.22

 regard

0.21

 indifferent

0.21

æ¯«

0.21

 cared

0.20

 cares

0.19

 lack

0.19

Activations Density 0.163%