INDEX

Explanations

punctuation and closing quotes

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

AXB

-0.10

YRO

-0.10

trys

-0.10

Tut

-0.09

 Dare

-0.09

 newcom

-0.09

oli

-0.08

Ary

-0.08

otu

-0.08

 ########.

-0.08

POSITIVE LOGITS

IGGER

0.09

 preferredStyle

0.08

trap

0.08

 Synthetic

0.08

ftware

0.08

Chr

0.08

_cred

0.08

 Scarlet

0.08

Signed

0.08

 "}\n

0.08

Activations Density 0.174%

No Known Activations