INDEX

Explanations

references to and discussion about books

oai_token-act-pair · gpt-3.5-turbo

references to books

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

00200000

-0.77

ALTH

-0.69

 Normandy

-0.68

isin

-0.67

Yin

-0.66

 Territories

-0.66

 Security

-0.63

 dism

-0.62

 Pearce

-0.61

 Hannity

-0.59

POSITIVE LOGITS

stores

1.34

 books

1.22

books

1.18

marks

1.10

worms

1.05

book

1.03

marked

0.94

wal

0.93

banks

0.93

poons

0.93

Activations Density 0.015%

references to and discussion about books

references to books

No Comments

No Known Activations