INDEX

Explanations

phrases related to events or organizations with the acronym "SW"

oai_token-act-pair · gpt-3.5-turbo

references to a specific location or identifier, likely related to a geographic region or organization

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 dÃ©

-0.74

 fluor

-0.68

 Gene

-0.67

 Prohibition

-0.66

 carcin

-0.64

 patri

-0.64

 Johnny

-0.63

cake

-0.62

 tumor

-0.62

 supposed

-0.62

POSITIVE LOGITS

SW

4.58

SW

2.11

sw

1.74

NW

1.62

Sw

1.42

CW

1.40

TW

1.35

SF

1.32

SD

1.27

RW

1.23

Activations Density 0.013%

phrases related to events or organizations with the acronym "SW"

references to a specific location or identifier, likely related to a geographic region or organization

No Comments

No Known Activations