INDEX

Explanations

valid python or JSON commands

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

-0.08

 proport

-0.08

 (...)

-0.07

 suce

-0.07

UC

-0.07

ifice

-0.07

los

-0.07

 Portug

-0.07

 guarante

-0.07

POSITIVE LOGITS

ÂĢÂĢ

0.13

 -*-č\n

0.12

Â¿ÃĤ

0.10

.Formatter

0.09

<|begin_of_text|>

0.09

 DÃ¼n

0.08

Âģ@

0.08

¦æĥħ

0.08

Ľi

0.08

******č\n

0.08

Activations Density 0.312%