INDEX

Explanations

completion or finality

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

entifier

-0.10

-translate

-0.09

irus

-0.09

Abd

-0.09

 writings

-0.09

 Conrad

-0.09

ï½Ľ

-0.08

 Spicer

-0.08

Abr

-0.08

tÄĽÅ¾

-0.08

POSITIVE LOGITS

 done

0.12

done

0.11

imen

0.10

 Done

0.09

RIEND

0.09

amo

0.09

ongan

0.09

owe

0.08

-done

0.08

Done

0.08

Activations Density 0.527%