INDEX

Explanations

quotations or references marked by "Qu" or "qu"

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

 Painter

-0.68

 Lyme

-0.67

 reconciliation

-0.67

fulness

-0.66

croft

-0.65

 mercury

-0.64

AGES

-0.64

 testosterone

-0.64

 disparate

-0.63

strap

-0.62

POSITIVE LOGITS

oted

1.25

oting

1.25

arantine

1.23

otation

1.21

artz

1.20

arter

1.17

adr

1.16

arters

1.14

atern

1.08

arant

1.07

Activations Density 0.019%

quotations or references marked by "Qu" or "qu"

No Comments

No Known Activations