INDEX

Explanations

short statements or headlines focusing on specific topics

oai_token-act-pair · gpt-3.5-turbo

specific phrases or concepts related to opinions or assessments

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

nor

-0.73

ades

-0.72

 eurozone

-0.71

escape

-0.66

alysed

-0.66

 nowhere

-0.65

 intervened

-0.64

 escaping

-0.62

ammed

-0.62

 Buch

-0.60

POSITIVE LOGITS

Firstly

0.92

Overview

0.89

First

0.79

Appearance

0.78

 Firstly

0.77

Name

0.77

ccording

0.76

Reason

0.76

Examples

0.75

Introduction

0.75

Activations Density 0.415%

short statements or headlines focusing on specific topics

specific phrases or concepts related to opinions or assessments

No Comments

No Known Activations