INDEX

Explanations

phrases related to specific names, particularly "Andrews" and "Pattison"

oai_token-act-pair · gpt-3.5-turbo

the name "Andrews" prominently referenced in the document

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

ounty

-0.81

ozyg

-0.75

itably

-0.74

orescent

-0.74

atorial

-0.74

onest

-0.73

acts

-0.68

ASED

-0.66

naire

-0.66

crime

-0.66

POSITIVE LOGITS

 Andrews

1.22

shire

0.91

endor

0.80

sburg

0.78

boro

0.77

kinson

0.74

butt

0.74

AFB

0.73

 Lans

0.73

combe

0.73

Activations Density 0.007%

phrases related to specific names, particularly "Andrews" and "Pattison"

the name "Andrews" prominently referenced in the document

No Comments

No Known Activations