INDEX

Explanations

programming constructs

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

.printStackTrace

-0.12

oreferrer

-0.11

 Ø¨ÙĪØ§Ø¨Ø©

-0.11

endregion

-0.10

olumn

-0.09

jspx

-0.09

 Boeh

-0.09

fty

-0.08

POSITIVE LOGITS

hape

0.09

DIST

0.09

 arsenal

0.09

ngr

0.09

 Giang

0.09

igon

0.09

 Picasso

0.08

ServletContext

0.08

illis

0.08

qid

0.08

Activations Density 0.009%