INDEX

Explanations

organizations or individuals facing scrutiny, criticism, or investigation

oai_token-act-pair · gpt-3.5-turbo

phrases indicating being "under" some form of scrutiny or pressure

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

HK

-0.72

DragonMagazine

-0.69

 Scand

-0.64

 VIDEOS

-0.63

STL

-0.63

Tur

-0.63

Hello

-0.62

ILLE

-0.61

 Tucson

-0.60

 Begins

-0.60

POSITIVE LOGITS

dogs

1.05

pins

1.02

neath

0.96

lined

0.95

graduate

0.95

covered

0.90

whelming

0.90

lining

0.89

cutting

0.89

lying

0.88

Activations Density 0.041%

organizations or individuals facing scrutiny, criticism, or investigation

phrases indicating being "under" some form of scrutiny or pressure

No Comments

No Known Activations