INDEX

Explanations

organized information, potentially in a list format

oai_token-act-pair · gpt-3.5-turbo

disconnected expressions or phrases that may highlight various topics or sentiments

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

 imperson

-0.85

 estranged

-0.75

 undermin

-0.71

 dissu

-0.70

 plur

-0.69

 interpre

-0.69

 intervened

-0.69

 antid

-0.67

 divest

-0.67

hed

-0.67

POSITIVE LOGITS

Posted

1.54

posted

1.32

Advertisements

1.28

ÂłÂł

1.26

ÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂł

1.25

Anyway

1.24

ÂłÂłÂł

1.21

Summary

1.21

Introduction

1.19

ÂłÂłÂłÂłÂłÂłÂłÂł

1.19

Activations Density 0.322%

organized information, potentially in a list format

disconnected expressions or phrases that may highlight various topics or sentiments

No Comments

No Known Activations