INDEX

Explanations

instances of quantitative information, such as percentages or numbers

oai_token-act-pair · gpt-3.5-turbo

phrases related to statistical data or reports on social issues

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

ĸļ

-0.81

xual

-0.77

llah

-0.77

abilia

-0.73

bek

-0.72

aper

-0.71

erker

-0.71

hemer

-0.70

 glim

-0.70

sung

-0.69

POSITIVE LOGITS

 Guam

1.52

 Uruguay

1.37

 Switzerland

1.36

 Rhode

1.36

 Tasmania

1.36

 Mexico

1.34

Mexico

1.33

 Chile

1.33

 Nebraska

1.33

 Hawaii

1.33

Activations Density 0.490%

instances of quantitative information, such as percentages or numbers

phrases related to statistical data or reports on social issues

No Comments

No Known Activations