INDEX

Explanations

addresses listed with state abbreviations at the end

oai_token-act-pair · gpt-3.5-turbo

U.S. state abbreviations and location identifiers

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

ufact

-0.78

lihood

-0.76

-0.75

ãĤ¨ãĥ«

-0.74

bats

-0.73

toggle

-0.72

é¾įå

-0.71

mentioned

-0.68

ãĤ»

-0.68

months

-0.66

POSITIVE LOGITS

 Supervisor

0.79

 Dept

0.76

ARA

0.71

JR

0.67

LEY

0.66

 STATE

0.64

 Admission

0.64

WA

0.63

 Examiner

0.63

RA

0.62

Activations Density 0.029%

addresses listed with state abbreviations at the end

U.S. state abbreviations and location identifiers

No Comments

No Known Activations