INDEX

Explanations

references to governmental and official entities, particularly offices

oai_token-act-pair · gpt-3.5-turbo

references to various government and organizational offices

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

mos

-0.74

 legends

-0.68

 cann

-0.64

 toget

-0.64

ante

-0.64

bass

-0.64

 compat

-0.64

amaz

-0.64

otaur

-0.63

 induct

-0.62

POSITIVE LOGITS

 Office

3.84

Office

3.06

office

2.44

 office

2.36

 offices

1.77

 Department

1.49

 Bureau

1.43

 Dept

1.32

 Agency

1.29

 Directorate

1.26

Activations Density 0.011%

references to governmental and official entities, particularly offices

references to various government and organizational offices

No Comments

No Known Activations