INDEX

Explanations

characteristics and properties

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

adol

-0.09

ìŀ¡

-0.09

Sag

-0.09

duino

-0.09

ropolis

-0.09

.userInteractionEnabled

-0.09

etherlands

-0.08

_Lean

-0.08

éĺµ

-0.08

anch

-0.08

POSITIVE LOGITS

 contents

0.15

 properties

0.15

contents

0.13

 characteristics

0.12

 capabilities

0.12

properties

0.12

 attributes

0.12

 behavior

0.11

 Properties

0.11

capabilities

0.11

Activations Density 0.159%