INDEX

Explanations

identifying names and definitions in 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

Not in Any Lists

No Comments

Negative Logits

“;ê·¸

-0.13

_Tis

-0.12

Intialized

-0.12

_Lean

-0.11

CallCheck

-0.09

Ð±ÑĥÐ´ÑĮ

-0.09

"),"

-0.09

@nate

-0.09

íĨłíĨł

-0.09

_Osc

-0.09

POSITIVE LOGITS

ÂĢÂ

0.21

ÂĿ

0.13

the

0.11

Âĺ

0.11

Ë

0.10

","","

0.10

ÂĢÂĻ

0.10

gnore

0.09

[]"

0.09

ropri

0.08

Activations Density 0.205%