INDEX

Explanations

sections or headings in technical or instructional documents

oai_token-act-pair · gpt-3.5-turbo

structured content in informational articles

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.8.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 anytime

-0.77

 Carlton

-0.69

 tweeted

-0.68

 blond

-0.67

 twent

-0.64

 actress

-0.63

 dudes

-0.63

 champ

-0.63

 dude

-0.63

 guiActive

-0.63

POSITIVE LOGITS

 Conclusion

1.28

 Definitions

1.26

 Introduction

1.20

 Contents

1.14

Overview

1.12

 Overview

1.11

Methods

1.09

Introduction

1.07

 Discussion

1.07

 Purpose

1.05

Activations Density 0.249%

sections or headings in technical or instructional documents

structured content in informational articles

No Comments

No Known Activations