INDEX

Explanations

instances of the word "dollars" followed by a number indicating a large amount of money

oai_token-act-pair · gpt-3.5-turbo

references to large sums of money

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ICLE

-0.76

 Stru

-0.72

 Nare

-0.68

 Prov

-0.66

 Blaz

-0.64

âĹ¼

-0.64

Maw

-0.63

Ku

-0.63

 Halls

-0.62

 Gors

-0.62

POSITIVE LOGITS

 dollars

1.18

cci

0.99

$$

0.97

 laundering

0.96

dollar

0.96

aundering

0.95

/$

0.91

($

0.89

 worth

0.88

USD

0.87

Activations Density 0.023%

instances of the word "dollars" followed by a number indicating a large amount of money

references to large sums of money

No Comments

No Known Activations