INDEX

Explanations

codes, numbers, and hashtags

oai_token-act-pair · gpt-3.5-turbo

references to social media or online content elements

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

 dexter

-0.81

 Franch

-0.80

alach

-0.79

 Brav

-0.79

æ³

-0.73

Mach

-0.72

wana

-0.71

 Minotaur

-0.71

BIL

-0.70

rir

-0.70

POSITIVE LOGITS

 transcript

0.94

 Photograph

0.89

 Photo

0.89

Text

0.88

 photograph

0.88

 Newspaper

0.88

picture

0.85

 screenshot

0.83

 diary

0.83

 Videos

0.83

Activations Density 0.452%

codes, numbers, and hashtags

references to social media or online content elements

No Comments

No Known Activations