INDEX

Explanations

error message you're seeing

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

Negative Logits

Gam

-0.10

mÃ©na

-0.10

ÙĪØ§Ø¡

-0.10

çĶĺ

-0.09

yps

-0.09

SIL

-0.09

 Ð³Ð¾Ð»

-0.08

Iso

-0.08

raz

-0.08

 Ð¿Ð¾Ð±Ð°

-0.08

POSITIVE LOGITS

 indicates

0.15

 suggests

0.15

 indicate

0.15

 suggest

0.14

 Indicates

0.12

suggest

0.12

 youre

0.12

 clue

0.12

 indication

0.11

sug

0.11

Activations Density 0.026%

No Known Activations