INDEX

Explanations

python list comprehensions

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

 Engel

-0.10

::.

-0.10

 Binary

-0.09

 Randolph

-0.09

:.:

-0.09

 binary

-0.09

irie

-0.09

Ã¥

-0.09

 Primitive

-0.09

oppers

-0.09

POSITIVE LOGITS

 comprehension

0.25

 compreh

0.18

 filter

0.18

 lambda

0.16

filter

0.15

zip

0.13

 expression

0.13

ilter

0.13

 comprend

0.13

 Filter

0.13

Activations Density 0.083%