INDEX

Explanations

the word "module" or a very close variant

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

 Braw

-0.77

 Bellow

-0.73

 Belo

-0.71

 Daria

-0.69

 Eugenia

-0.68

 Majefty

-0.68

 Marten

-0.68

Mâ

-0.68

 Sarmiento

-0.66

MAV

-0.66

POSITIVE LOGITS

 modules

2.22

 Modules

2.19

 Module

2.05

 module

2.00

modules

1.98

Modules

1.97

 MODULE

1.90

Module

1.89

module

1.85

MODULE

1.82

Activations Density 0.186%

No Known Activations