INDEX

Explanations

mentions of violent incidents or crimes related to warfare

oai_token-act-pair · gpt-3.5-turbo

references to violent events and fatalities associated with conflict or terrorism

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

 FANTASY

-0.99

Consumer

-0.98

Kids

-0.98

Madison

-0.89

 Consumers

-0.85

 brunch

-0.85

 Consumer

-0.84

opia

-0.83

Seattle

-0.81

consumer

-0.79

POSITIVE LOGITS

 assassinated

1.50

 Gaddafi

1.37

 detained

1.27

 jailed

1.25

 Islamist

1.22

 martyr

1.19

 assassinate

1.19

 exiled

1.19

 militant

1.18

 Kazakh

1.17

Activations Density 0.801%

mentions of violent incidents or crimes related to warfare

references to violent events and fatalities associated with conflict or terrorism

No Comments

No Known Activations