INDEX

Explanations

target-related words or phrases

oai_token-act-pair · gpt-3.5-turbo

instances of the word "target" and its variations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

loo

-0.78

maid

-0.73

iHUD

-0.70

fo

-0.70

SourceFile

-0.70

lycer

-0.68

cia

-0.66

 Expedition

-0.64

ublic

-0.64

BuyableInstoreAndOnline

-0.62

POSITIVE LOGITS

ted

1.37

ting

0.99

ched

0.83

 audience

0.75

 squarely

0.72

eers

0.71

oided

0.70

ivated

0.69

 targets

0.67

chers

0.66

Activations Density 0.046%

target-related words or phrases

instances of the word "target" and its variations

No Comments

No Known Activations