INDEX

Explanations

prone to breaking

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

 inconsistency

-0.13

oka

-0.11

-0.10

çĽĳåĲ¬é¡µéĿ¢

-0.10

 inconsistencies

-0.09

/Error

-0.09

 Fault

-0.09

Configurer

-0.09

asio

-0.08

nad

-0.08

POSITIVE LOGITS

 Curse

0.15

 multic

0.14

 curse

0.14

 coll

0.13

 brit

0.12

deg

0.12

 curses

0.11

alias

0.11

 bias

0.11

 alias

0.11

Activations Density 0.080%