INDEX

Explanations

references to computer programming concepts and tools

oai_token-act-pair · gpt-3.5-turbo

references to computer programs and programming concepts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Accessory

-0.69

ModLoader

-0.68

 stake

-0.66

 Pradesh

-0.65

xious

-0.62

 Tall

-0.62

 exting

-0.61

 Marketable

-0.61

?????-

-0.60

 Blackwell

-0.60

POSITIVE LOGITS

mable

1.83

matically

1.45

matic

1.45

mers

1.21

atically

1.04

zee

0.99

matical

0.99

mercial

0.96

 Gutenberg

0.94

manship

0.93

Activations Density 0.025%

references to computer programming concepts and tools

references to computer programs and programming concepts

No Comments

No Known Activations