INDEX

Explanations

identify, discover, generate

the word "to" when it introduces an infinitive verb or purpose clause.

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

Nad

-0.11

 knowing

-0.10

 knew

-0.09

 Knowing

-0.09

SCP

-0.09

Sources

-0.09

edad

-0.09

ì¦Ŀ

-0.09

 overt

-0.08

Tay

-0.08

POSITIVE LOGITS

 identified

0.22

 identify

0.21

 spot

0.19

åıĳçİ°

0.19

 patterns

0.19

 discover

0.19

 pattern

0.18

 find

0.18

 identifies

0.18

 finding

0.17

Activations Density 0.131%