INDEX

Explanations

technical terms or phrases related to additional or extra information

oai_token-act-pair · gpt-3.5-turbo

terms related to supplementary or technical content

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

uph

-0.85

azz

-0.76

Sov

-0.69

Roc

-0.69

Monster

-0.68

 pessim

-0.67

 Mafia

-0.67

 Bohem

-0.66

doms

-0.65

bour

-0.65

POSITIVE LOGITS

 Supplemental

2.91

 supplemental

2.41

 Supplementary

2.30

 supplementary

2.03

Supplement

1.83

SI

1.60

 Appendix

1.40

 appendix

1.32

 annexed

1.23

 Applied

1.22

Activations Density 0.054%

technical terms or phrases related to additional or extra information

terms related to supplementary or technical content

No Comments

No Known Activations