INDEX

Explanations

function return code or checking condition

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

 marsh

-0.12

 Gallagher

-0.09

 Prem

-0.09

 butt

-0.09

 getRequest

-0.09

iss

-0.08

 gall

-0.08

kle

-0.08

ang

-0.08

POSITIVE LOGITS

 handle

0.12

 check

0.10

 process

0.10

cdecl

0.10

 HANDLE

0.10

(get

0.10

callee

0.10

 helper

0.09

crate

0.09

,ï½¤

0.09

Activations Density 0.074%