INDEX

Explanations

numbers and technical codes in the text

oai_token-act-pair · gpt-3.5-turbo

specific numerical identifiers or codes related to products or studies

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

querque

-0.73

enegger

-0.68

oooooooooooooooo

-0.66

ogene

-0.66

 snowball

-0.63

 unemployed

-0.62

 bere

-0.62

 stroke

-0.61

inations

-0.60

 Oath

-0.60

POSITIVE LOGITS

DM

0.98

XX

0.97

CAP

0.89

VB

0.88

CB

0.87

GW

0.85

XXX

0.84

eus

0.84

ZE

0.84

UF

0.83

Activations Density 0.090%

numbers and technical codes in the text

specific numerical identifiers or codes related to products or studies

No Comments

No Known Activations