INDEX

Explanations

images or image-related HTML code

oai_token-act-pair · gpt-3.5-turbo

occurrences of image-related content

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

 termination

-0.80

 termin

-0.78

 parity

-0.70

 choke

-0.68

itive

-0.66

ACTED

-0.64

 Sawyer

-0.63

 throat

-0.63

 paraly

-0.63

 backdoor

-0.63

POSITIVE LOGITS

img

3.74

images

2.64

image

1.79

cdn

1.77

uploads

1.62

img

1.57

static

1.40

boards

1.38

files

1.29

forums

1.25

Activations Density 0.029%

images or image-related HTML code

occurrences of image-related content

No Comments

No Known Activations