INDEX

Explanations

phrases containing arrows or direction symbols

oai_token-act-pair · gpt-3.5-turbo

symbols or characters that are non-standard or potentially censored

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

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

 photoc

-0.72

 transc

-0.71

 interf

-0.70

 synd

-0.70

 cloaked

-0.67

 smoot

-0.67

bes

-0.66

 mete

-0.65

 telev

-0.64

 cumbers

-0.64

POSITIVE LOGITS

Ļ

1.91

¬

1.30

ħ

1.29

į

1.28

Ĵ

1.26

ĺ

1.25

«

1.24

ª

1.21

Ĺ

1.20

¤

1.19

Activations Density 0.388%

phrases containing arrows or direction symbols

symbols or characters that are non-standard or potentially censored

No Comments

No Known Activations