INDEX

Explanations

references to the University of Wisconsin (UW)

oai_token-act-pair · gpt-3.5-turbo

mentions of academic institutions or universities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 Staten

-0.95

 Liberty

-0.81

 Virgin

-0.80

 Southern

-0.79

Dick

-0.75

lim

-0.74

 Harlem

-0.74

 Kelly

-0.74

 Notting

-0.73

 Atlantic

-0.71

POSITIVE LOGITS

UW

3.06

UA

1.59

UNC

1.38

 Petersen

1.36

UR

1.30

WB

1.24

PLA

1.21

BYU

1.20

tut

1.19

Tut

1.18

Activations Density 0.051%

references to the University of Wisconsin (UW)

mentions of academic institutions or universities

No Comments

No Known Activations