INDEX

Explanations

words related to starting, organizing, and managing various tasks or projects

oai_token-act-pair · gpt-3.5-turbo

the word "start" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

 Seym

-0.79

 Vaugh

-0.73

 Moroc

-0.69

 helicop

-0.68

 Niet

-0.66

enegger

-0.64

 vulner

-0.64

edIn

-0.63

emale

-0.63

ayson

-0.61

POSITIVE LOGITS

 largeDownload

0.69

\":

0.57

 debian

0.50

 Disc

0.50

 JUSTICE

0.50

 GeForce

0.49

 Xperia

0.48

 heats

0.47

ARM

0.46

 Isis

0.46

Activations Density 0.915%

words related to starting, organizing, and managing various tasks or projects

the word "start" in various contexts

No Comments

No Known Activations