INDEX

Explanations

section headings

This neuron detects when a sentence begins with the word “In.”

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

 gördüğünüz

0.75

 veldig

0.73

 máte

0.72

 majest

0.70

 darle

0.67

 pesky

0.67

 perfek

0.67

 bikin

0.66

 banget

0.66

 yako

0.66

POSITIVE LOGITS

Keywords

1.82

 Keywords

1.73

ABSTRACT

1.53

 INTRODUCTION

1.52

INTRODUCTION

1.50

 ABSTRACT

1.50

KEYWORDS

1.42

Abstract

1.40

 Abstract

1.36

Introduction

1.30

Activations Density 0.713%