INDEX

Explanations

knowledge of specific subjects

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

 prelim

-0.09

ylon

-0.09

Fem

-0.09

ugh

-0.09

bia

-0.08

æĻº

-0.08

à¸ªà¸²à¸£

-0.08

âĶĲ

-0.08

Baz

-0.08

ogh

-0.08

POSITIVE LOGITS

how

0.16

å¦Ĥä½ķ

0.13

 basic

0.11

 Concepts

0.11

 concepts

0.11

ingly

0.10

how

0.10

lox

0.10

 both

0.10

 principles

0.09

Activations Density 0.053%