INDEX

Explanations

phrases related to work or professional activities

oai_token-act-pair · gpt-3.5-turbo

references to someone's work or contributions

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Ukrain

-0.97

wcs

-0.70

 Flavoring

-0.68

anamo

-0.67

 Rohing

-0.66

 Predators

-0.66

 constitu

-0.66

iren

-0.65

 Diet

-0.64

EStream

-0.63

POSITIVE LOGITS

station

1.32

 ethic

1.31

flows

1.25

manship

1.16

bench

1.15

fare

1.11

aday

1.07

work

0.99

horse

0.97

ername

0.95

Activations Density 0.063%

phrases related to work or professional activities

references to someone's work or contributions

No Comments

No Known Activations