INDEX

Explanations

theory of everything

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

 spac

-0.10

è¿Ķ

-0.09

trl

-0.08

 astronomers

-0.08

 lifted

-0.08

Dun

-0.08

wand

-0.08

CIS

-0.08

ammad

-0.08

YPE

-0.08

POSITIVE LOGITS

 Loop

0.18

 string

0.17

 String

0.17

 inflation

0.16

 loop

0.15

Loop

0.15

String

0.14

/String

0.14

 theories

0.13

/string

0.13

Activations Density 0.074%