INDEX

Explanations

code punctuation and keywords

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

gir

-0.10

imp

-0.10

 Pompe

-0.09

ecome

-0.09

pup

-0.09

utherland

-0.09

ÂłÂłÂłÂłÂł

-0.09

 shelter

-0.09

imp

-0.09

 minor

-0.09

POSITIVE LOGITS

0.13

ãĢĢãĢĢãĢĢ ãĢĢ

0.11

 ãĢĢ ãĢĢ ãĢĢ

0.11

IFS

0.10

ãĢĢ ãĢĢ ãĢĢ

0.09

.habbo

0.09

ursal

0.09

\t\t

0.09

ãĢĢãĢĢãĢĢãĢĢãĢĢ

0.09

Activations Density 0.018%