INDEX

Explanations

references to the word "content."

oai_token-act-pair · gpt-3.5-turbo

references to digital content

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

rolet

-0.75

Ram

-0.71

 Seasons

-0.70

 Rasmussen

-0.69

STRUCT

-0.64

 Friendship

-0.62

 damn

-0.62

 Satellite

-0.60

 Thornton

-0.60

 Hamm

-0.60

POSITIVE LOGITS

 content

1.30

edly

1.28

Content

0.98

content

0.93

 Content

0.87

estine

0.81

cms

0.76

 contents

0.75

area

0.75

leground

0.74

Activations Density 0.014%

references to the word "content."

references to digital content

No Comments

No Known Activations