INDEX

Explanations

references to blog posts

oai_token-act-pair · gpt-3.5-turbo

references to blogs and blog posts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

inho

-0.76

BuyableInstoreAndOnline

-0.75

emale

-0.73

 Liberties

-0.70

 Franch

-0.64

 Shinra

-0.64

ZI

-0.64

 Lauder

-0.63

odka

-0.63

NESS

-0.62

POSITIVE LOGITS

osphere

1.15

gers

0.97

post

0.97

 postings

0.91

blogs

0.91

 blog

0.89

blog

0.88

 posts

0.88

posts

0.88

ged

0.86

Activations Density 0.014%

references to blog posts

references to blogs and blog posts

No Comments

No Known Activations