INDEX

Explanations

mentions of specific religious denominations or churches

oai_token-act-pair · gpt-3.5-turbo

references to specific Christian denominations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

rg

-0.83

)=(

-0.81

kok

-0.80

coat

-0.75

lessly

-0.74

dest

-0.74

agame

-0.72

 Rove

-0.72

bg

-0.70

grab

-0.70

POSITIVE LOGITS

 Episcopal

1.29

 Methodist

1.16

 Church

1.07

 Presbyterian

1.00

 church

0.97

 clergy

0.93

 Lutheran

0.92

 communion

0.89

 missionaries

0.89

 missionary

0.87

Activations Density 0.017%

mentions of specific religious denominations or churches

references to specific Christian denominations

No Comments

No Known Activations