INDEX

Explanations

phrases related to government officials and government agencies

oai_token-act-pair · gpt-3.5-turbo

significant governmental positions and their appointees

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

 neighbourhood

-0.83

 neighbourhoods

-0.82

 Mandela

-0.79

 Cardiff

-0.79

 footballer

-0.78

qv

-0.75

 toget

-0.74

 Catalan

-0.74

 boxing

-0.73

 poss

-0.73

POSITIVE LOGITS

 agency

1.79

HHS

1.62

DOE

1.46

agency

1.44

 USDA

1.44

EPA

1.43

 Agency

1.42

 agencies

1.40

DHS

1.39

 Administrator

1.38

Activations Density 0.653%

phrases related to government officials and government agencies

significant governmental positions and their appointees

No Comments

No Known Activations