INDEX

Explanations

references to a specific word or concept, 'contra.'

oai_token-act-pair · gpt-3.5-turbo

mentions of concepts related to defiance or violation of rules and regulations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

ļéĨĴ

-0.80

 Assass

-0.79

onym

-0.74

 Nanto

-0.71

FSA

-0.70

Rated

-0.69

istics

-0.67

 Nationwide

-0.67

NetMessage

-0.63

 Columbus

-0.63

POSITIVE LOGITS

 contra

1.36

ptions

1.10

ption

0.98

ventions

0.79

ven

0.78

emonium

0.76

xon

0.76

asca

0.76

vention

0.75

etheless

0.72

Activations Density 0.008%

references to a specific word or concept, 'contra.'

mentions of concepts related to defiance or violation of rules and regulations

No Comments

No Known Activations