INDEX

Explanations

keywords related to news articles

oai_token-act-pair · gpt-3.5-turbo

instances of the word "MORE"

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

stood

-0.84

ades

-0.76

ted

-0.76

med

-0.73

aper

-0.73

etime

-0.72

arist

-0.69

riber

-0.69

eb

-0.69

ibel

-0.68

POSITIVE LOGITS

MORE

1.38

agascar

0.87

 convol

0.86

 than

0.82

quickShipAvailable

0.81

 IMAGES

0.81

HEAD

0.81

ado

0.80

rontal

0.79

FTWARE

0.78

Activations Density 0.006%

keywords related to news articles

instances of the word "MORE"

No Comments

No Known Activations