INDEX

Explanations

intros or introductions within texts

oai_token-act-pair · gpt-3.5-turbo

instances of the word "intro" and related variations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ware

-0.77

land

-0.70

 parcels

-0.69

 munitions

-0.69

endants

-0.68

 Maritime

-0.68

wald

-0.66

 Century

-0.66

 mills

-0.65

UST

-0.65

POSITIVE LOGITS

 intro

3.65

 Intro

2.80

 introductory

1.36

 introdu

1.22

 Introduction

1.20

 introduction

1.15

 undergrad

1.07

 opener

1.06

 autobi

0.97

 emot

0.95

Activations Density 0.023%

intros or introductions within texts

instances of the word "intro" and related variations

No Comments

No Known Activations