INDEX

Explanations

numbers, especially those related to quantitative figures and measurements

oai_token-act-pair · gpt-3.5-turbo

numeric values and references to quantities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

atta

-0.71

onductor

-0.69

atown

-0.68

berus

-0.68

ktop

-0.66

activate

-0.66

mble

-0.65

roup

-0.65

£ı

-0.63

ivation

-0.62

POSITIVE LOGITS

 bucks

1.26

 cents

1.20

 billion

1.15

 trillion

1.11

 goddamn

1.09

 seconds

1.07

 dudes

1.07

 million

1.05

 years

1.01

 peas

0.99

Activations Density 0.219%

numbers, especially those related to quantitative figures and measurements

numeric values and references to quantities

No Comments

No Known Activations