INDEX

Explanations

keywords related to customization and settings

oai_token-act-pair · gpt-3.5-turbo

actions related to configuration, customization, or setting options

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

é¾įå¥ĳå£«

-0.80

BuyableInstoreAndOnline

-0.71

rium

-0.66

Defense

-0.64

eded

-0.64

ĸļ

-0.64

 endeavor

-0.63

bian

-0.60

Ben

-0.59

Lenin

-0.58

POSITIVE LOGITS

 yourself

0.99

ivably

0.90

 yourselves

0.88

 manually

0.82

 unlimited

0.80

 easily

0.80

 your

0.75

 freely

0.75

 anonymously

0.72

 anywhere

0.72

Activations Density 0.352%

keywords related to customization and settings

actions related to configuration, customization, or setting options

No Comments

No Known Activations