INDEX

Explanations

news headlines or article titles

oai_token-act-pair · gpt-3.5-turbo

calls to action or prompts for further reading

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

raf

-0.74

phia

-0.69

stood

-0.64

bats

-0.64

ffen

-0.62

 Belt

-0.62

rays

-0.61

unction

-0.61

angers

-0.59

aten

-0.58

POSITIVE LOGITS

 STOR

1.39

 STORY

1.36

 ALSO

1.34

OUR

1.29

 INFORMATION

1.28

 VIDEO

1.27

 APPLIC

1.26

 VIDEOS

1.24

 BELOW

1.22

 REPORT

1.21

Activations Density 0.041%

news headlines or article titles

calls to action or prompts for further reading

No Comments

No Known Activations