INDEX

Explanations

bear in the forest

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

 ...\n\n\n\n

-0.09

-0.08

-0.07

for

-0.07

zb

-0.07

qu

-0.06

.scalablytyped

-0.06

 Direct

-0.06

ystick

-0.06

POSITIVE LOGITS

¶Į

0.13

******č\n

0.13

¦æĥħ

0.12

ÂĢÂĢ

0.11

.Formatter

0.11

įng

0.10

 -*-č\n

0.10

Â¿ÃĤ

0.10

<|begin_of_text|>

0.09

ÂĤÃĮ

0.09

Activations Density 0.696%