INDEX

Explanations

internal forward directive

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 détaillée

-0.71

endal

-0.71

크

-0.70

فره

-0.69

ztus

-0.69

Zus

-0.67

hagy

-0.67

Lexus

-0.66

ğum

-0.66

reak

-0.66

POSITIVE LOGITS

 forward

1.28

 forwarding

1.27

 dispatcher

1.26

dispatcher

1.22

Dispatcher

1.13

Dispatch

1.12

 dispatch

1.10

forward

1.09

 Dispatcher

1.09

转发

1.07

Activations Density 0.005%