INDEX

Explanations

ignore or disregard

The neuron is triggered by imperative directives—especially “ignore” or “pay no attention to” style commands telling the reader to dismiss something.

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

is

-1.79

除此之外

-1.35

-1.26

 like

-1.24

can

-1.23

 Consequently

-1.22

AND

-1.20

 where

-1.19

і

-1.18

 WITH

-1.16

POSITIVE LOGITS

จำนวน

1.35

 během

1.31

 provinciale

1.30

 titoli

1.28

に入り

1.27

曏

1.24

」、

1.23

decoración

1.21

ladung

1.21

 recetas

1.20

Activations Density 0.030%