INDEX

Explanations

references to news sources and citations within a text

oai_token-act-pair · gpt-3.5-turbo

sources and credits for images and media

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 reconsider

-0.65

 earnest

-0.62

 Clancy

-0.60

self

-0.58

 chapters

-0.58

 foundations

-0.58

ardy

-0.58

 elev

-0.57

 Crowley

-0.57

stood

-0.57

POSITIVE LOGITS

 IMAGES

1.33

reenshot

1.11

utterstock

1.09

UTERS

1.08

 PHOTO

1.07

photo

1.06

Photo

1.05

 Shutterstock

0.99

 Photo

0.99

images

0.98

Activations Density 0.262%

references to news sources and citations within a text

sources and credits for images and media

No Comments

No Known Activations