INDEX

Explanations

computer program

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

aku

-0.10

casting

-0.10

 bait

-0.09

iant

-0.09

 Nicholas

-0.08

 households

-0.08

stad

-0.08

Saw

-0.08

 Witt

-0.08

 adul

-0.08

POSITIVE LOGITS

ized

0.16

 program

0.13

ised

0.13

 entity

0.11

 software

0.11

 programa

0.11

 programme

0.11

IZED

0.11

AI

0.10

ç¨ĭåºı

0.10

Activations Density 0.029%