INDEX

Explanations

phrases or labels often associated with the end of text or the completion of a document

oai_token-act-pair · gpt-3.5-turbo

the end sequences of text or separators between distinct sections of information

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

 Azerb

-0.38

govtrack

-0.31

ĪĴ

-0.31

 drowned

-0.30

 newcom

-0.29

 miscon

-0.29

 millenn

-0.29

 Huss

-0.28

 pilgr

-0.28

 vandal

-0.28

POSITIVE LOGITS

BRE

0.38

Description

0.37

Share

0.35

AppData

0.35

 Shutterstock

0.34

wcsstore

0.34

Product

0.34

Loading

0.34

ebook

0.34

Abstract

0.33

Activations Density 0.783%

phrases or labels often associated with the end of text or the completion of a document

the end sequences of text or separators between distinct sections of information

No Comments

No Known Activations