INDEX

Explanations

Without looking

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 ebenfalls

1.12

工作室

1.12

 Washington

1.09

 Honor

1.08

 アルミ

1.08

 Miami

1.07

 Austin

1.06

 Morris

1.05

 Mosby

1.05

 Memorial

1.03

POSITIVE LOGITS

ными

1.09

foresaid

1.06

Các

1.00

Rightarrow

0.98

relevant

0.98

 موجب

0.95

ల

0.95

predictor

0.95

ົດ

0.94

PRINC

0.92

Activations Density 0.126%