INDEX

Explanations

references to specific custom terms or phrases

oai_token-act-pair · gpt-3.5-turbo

terms related to employment, work, and customizable content

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

olkien

-0.70

Downloadha

-0.67

 Scotia

-0.66

agnar

-0.66

DonaldTrump

-0.65

 reper

-0.63

EGA

-0.62

anners

-0.62

SHE

-0.62

 mosqu

-0.61

POSITIVE LOGITS

er

2.44

ers

1.94

ership

1.84

erate

1.41

erd

1.31

erness

1.30

eri

1.29

ed

1.26

ered

1.22

ation

1.22

Activations Density 0.088%

references to specific custom terms or phrases

terms related to employment, work, and customizable content

No Comments

No Known Activations