INDEX

Explanations

phrases related to planning and carrying out harmful actions or schemes

oai_token-act-pair · gpt-3.5-turbo

phrases related to plotting or planning violent acts or crimes

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

framework

-0.73

imentary

-0.72

 ðŁĳ

-0.72

 Quartz

-0.70

required

-0.68

 partName

-0.67

HUD

-0.66

ä¸Ĭ

-0.65

facing

-0.65

reflect

-0.64

POSITIVE LOGITS

 terrorist

1.26

 murder

1.25

 murders

1.24

 terror

1.24

 mayhem

1.23

 terrorism

1.22

 massacres

1.20

 destruction

1.19

 attacks

1.19

 atrocities

1.19

Activations Density 0.324%

phrases related to planning and carrying out harmful actions or schemes

phrases related to plotting or planning violent acts or crimes

No Comments

No Known Activations