INDEX

Explanations

specific instances of the word "Dh" or "dh" followed by a number

oai_token-act-pair · gpt-3.5-turbo

references to the abbreviation "Dh" and its variations

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

 catalog

-0.65

omers

-0.65

 Blues

-0.64

 magnets

-0.63

Scot

-0.63

erman

-0.63

essee

-0.61

omer

-0.61

ous

-0.61

 helium

-0.61

POSITIVE LOGITS

Dh

4.16

dh

1.83

dh

1.63

 Bangl

1.59

 Bangladesh

1.50

Bh

1.37

DH

1.36

DH

1.30

 Dharma

1.28

Dj

1.26

Activations Density 0.017%

specific instances of the word "Dh" or "dh" followed by a number

references to the abbreviation "Dh" and its variations

No Comments

No Known Activations