INDEX

Explanations

who are you 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

errick

-0.09

 pride

-0.09

tainment

-0.09

ista

-0.09

 Shel

-0.09

 print

-0.09

 vari

-0.09

ï¾Ħ

-0.08

ocate

-0.08

anka

-0.08

POSITIVE LOGITS

ibri

0.10

 Bout

0.09

aho

0.09

ALES

0.09

 Baghd

0.09

 guys

0.09

 kidding

0.09

ponde

0.09

 datings

0.09

Activations Density 0.040%