INDEX

Explanations

structured data patterns typically found in computer programming, such as variable declarations and function definitions

oai_token-act-pair · gpt-3.5-turbo

structural elements in source code or data formats

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

 Hanson

-0.66

 mush

-0.62

 weaving

-0.61

 gymn

-0.60

 Yose

-0.59

 despair

-0.59

 bush

-0.59

bol

-0.59

 Aires

-0.59

 Reborn

-0.59

POSITIVE LOGITS

name

1.32

Name

1.28

uid

1.26

description

1.25

username

1.24

Description

1.22

href

1.19

 name

1.18

Filename

1.17

Version

1.16

Activations Density 0.252%

structured data patterns typically found in computer programming, such as variable declarations and function definitions

structural elements in source code or data formats

No Comments

No Known Activations