INDEX

Explanations

stack operations and balanced parentheses

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

OMB

-0.10

ouch

-0.09

ìĦ

-0.09

HBO

-0.09

 carc

-0.09

HB

-0.09

omb

-0.09

sup

-0.09

REE

-0.08

POSITIVE LOGITS

 stack

0.39

 Stack

0.36

æł

0.33

Stack

0.31

stack

0.30

 push

0.29

-stack

0.29

.Stack

0.29

(stack

0.28

.stack

0.28

Activations Density 0.047%