INDEX

Explanations

instructional or guiding language cues, such as "Let's", "So", and "First"

oai_token-act-pair · gpt-3.5-turbo

instances of introductory phrases or sentences

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

steen

-0.67

 ..."

-0.63

morrow

-0.63

realDonaldTrump

-0.62

 constitu

-0.61

TRUMP

-0.60

Romney

-0.59

 restores

-0.58

onement

-0.58

vernment

-0.57

POSITIVE LOGITS

 nutshell

0.93

 Concept

0.77

 Basics

0.75

Introduction

0.73

Overview

0.73

 Overview

0.72

 originally

0.71

Previous

0.71

 Designed

0.71

 Introduction

0.70

Activations Density 0.666%

instructional or guiding language cues, such as "Let's", "So", and "First"

instances of introductory phrases or sentences

No Comments

No Known Activations