INDEX

Explanations

dates and events mentioned within a document

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 tackling

-0.86

 pounding

-0.78

 combating

-0.76

 confronting

-0.76

 strangers

-0.75

 embracing

-0.73

 afar

-0.71

 debating

-0.71

 overwhelm

-0.70

 bashing

-0.70

POSITIVE LOGITS

 Citation

0.98

Contents

0.95

 NOTICE

0.89

SELECT

0.89

 Regist

0.85

license

0.84

RESULTS

0.84

 WARRANT

0.83

 ********************************

0.83

BuyableInstoreAndOnline

0.82

Activations Density 0.639%

dates and events mentioned within a document

No Comments

No Known Activations