INDEX

Explanations

allows or helps to

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

<|begin_of_text|>

-0.15

EMPLARY

-0.12

urette

-0.12

.printStackTrace

-0.10

TRGL

-0.09

ÂĢÂĢ

-0.09

BOOLE

-0.09

IIIK

-0.08

Â©Â©

-0.08

synthesize

-0.08

POSITIVE LOGITS

 help

0.15

 allow

0.13

 allows

0.12

 helps

0.11

 assist

0.11

help

0.11

åħģ

0.10

allow

0.10

let

0.10

 permet

0.09

Activations Density 0.162%