INDEX

Explanations

images or descriptions related to specific visual elements or characteristics

oai_token-act-pair · gpt-3.5-turbo

references to visual imagery or descriptive details related to banners, flags, and logos

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

reviewed

-0.91

Reviewer

-0.83

ife

-0.79

essen

-0.79

anship

-0.78

theless

-0.77

ibling

-0.77

Reviewed

-0.75

olicy

-0.72

estate

-0.71

POSITIVE LOGITS

 logo

1.64

 logos

1.48

 initials

1.47

 emblem

1.46

 slogan

1.43

 swast

1.42

 insign

1.42

 likeness

1.42

 inscription

1.39

 symbol

1.35

Activations Density 0.367%

images or descriptions related to specific visual elements or characteristics

references to visual imagery or descriptive details related to banners, flags, and logos

No Comments

No Known Activations