INDEX

Explanations

mentions of the country Australia

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "Australia."

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

outer

-0.77

ows

-0.76

Lens

-0.73

ppings

-0.73

lasses

-0.72

acco

-0.72

soDeliveryDate

-0.70

angler

-0.69

osen

-0.68

uler

-0.66

POSITIVE LOGITS

 Australia

0.97

Australia

0.89

 Aboriginal

0.88

 Australians

0.86

 Broadcasting

0.84

 Defence

0.84

gdala

0.82

 Samoa

0.78

 Geographic

0.77

apolis

0.77

Activations Density 0.027%

mentions of the country Australia

occurrences of the word "Australia."

No Comments

No Known Activations