INDEX

Explanations

references to locations in Australia, specifically New South Wales (NSW) and Queensland

oai_token-act-pair · gpt-3.5-turbo

mentions of regions in Australia, specifically New South Wales and Queensland

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

ãĥł

-0.78

bda

-0.77

aphael

-0.75

heimer

-0.73

ocobo

-0.71

thood

-0.70

cean

-0.70

arters

-0.69

toggle

-0.67

aci

-0.66

POSITIVE LOGITS

NSW

1.00

urst

0.83

FW

0.76

IRO

0.75

DEF

0.74

ADA

0.74

UC

0.74

RL

0.71

CC

0.70

 Defence

0.69

Activations Density 0.006%

references to locations in Australia, specifically New South Wales (NSW) and Queensland

mentions of regions in Australia, specifically New South Wales and Queensland

No Comments

No Known Activations