INDEX

Explanations

making actions

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

 making

-0.13

 makin

-0.10

ä½¿

-0.10

eru

-0.09

making

-0.09

[__

-0.09

 Making

-0.09

.intellij

-0.09

 Measures

-0.09

oya

-0.09

POSITIVE LOGITS

 happen

0.20

 rounds

0.13

/send

0.12

 strides

0.12

leine

0.11

appen

0.11

 mistakes

0.11

 progress

0.11

 Rounds

0.11

 available

0.11

Activations Density 0.128%