INDEX

Explanations

desired outcome or format

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

/devices

-0.09

ALLY

-0.09

Ã¼Ã§

-0.09

DCALL

-0.09

æı¡

-0.08

ally

-0.08

Haz

-0.08

âĢĮØ§ÙĨØ¨Ø§Ø±

-0.08

FixedSize

-0.08

 Loving

-0.08

POSITIVE LOGITS

/request

0.19

 outcome

0.18

/target

0.15

/app

0.15

 outcomes

0.14

/generated

0.14

Outcome

0.13

 amount

0.13

outcome

0.13

ably

0.12

Activations Density 0.020%