INDEX

Explanations

German dialogue and questions

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

Ã£y

-0.08

emand

-0.08

ARN

-0.08

ucher

-0.07

 &#8203;&#8203;

-0.07

ponce

-0.07

umbnails

-0.07

--[

-0.07

&amp

-0.07

Å¡

-0.07

POSITIVE LOGITS

 reluct

0.10

 modne

0.09

ÐĲÑĢÑħÑĸÐ²

0.09

NewProp

0.09

ëį°ìĿ´íĬ¸

0.09

Intialized

0.08

 safeg

0.08

AdapterManager

0.08

OffsetTable

0.08

>manual

0.08

Activations Density 0.072%