INDEX

Explanations

adjustments

The neuron detects occurrences of the word “adjust” (in any inflected form such as adjust, adjusted, adjusting, adjustment).

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

 Incluye

-1.80

Hyundai

-1.76

 hoher

-1.73

imantan

-1.73

 ответил

-1.70

But

-1.69

的一些

-1.67

sandalia

-1.65

σει

-1.64

 Añade

-1.63

POSITIVE LOGITS

The

2.19

谞

2.16

调整

2.03

 adjust

1.94

 bidra

1.88

adjust

1.86

𓐍

1.84

会儿

1.80

橚

1.80

 проблема

1.77

Activations Density 0.023%