INDEX

Explanations

diagrams and schematics

oai_token-act-pair · gpt-3.5-turbo

references to diagrams and schematics

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

ierra

-0.74

NetMessage

-0.69

eele

-0.67

 rainy

-0.66

ggies

-0.65

adies

-0.64

Ton

-0.63

ternity

-0.62

elope

-0.62

Pear

-0.61

POSITIVE LOGITS

matically

1.24

matic

1.24

matical

1.10

ming

1.03

 diagram

1.02

zag

0.97

 diagrams

0.96

mat

0.80

clip

0.78

eers

0.76

Activations Density 0.019%

diagrams and schematics

references to diagrams and schematics

No Comments

No Known Activations