INDEX

Explanations

phrases related to criminal activities and law enforcement

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

sections of text that prompt further reading or watching

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 12-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.hook_resid_post

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.11.hook_resid_post

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

aten

-0.76

 appe

-0.70

 manif

-0.70

 possession

-0.69

 orderly

-0.65

 spons

-0.64

 concentrated

-0.64

 imperson

-0.63

athered

-0.62

uthor

-0.62

POSITIVE LOGITS

 ALSO

1.47

 VIDEO

1.42

 STOR

1.39

 BELOW

1.38

 STORY

1.37

 INFORMATION

1.33

 VIDEOS

1.33

BOOK

1.31

 ARTICLE

1.28

 ABOUT

1.26

Activations Density 0.057%

phrases related to criminal activities and law enforcement

sections of text that prompt further reading or watching

No Comments

No Known Activations