INDEX

Explanations

mentions of violent actions such as assassination

oai_token-act-pair · gpt-3.5-turbo

references to assassination attempts and related discussions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

BuyableInstoreAndOnline

-0.82

Cola

-0.80

Sea

-0.78

Gam

-0.76

Va

-0.75

Bra

-0.75

Alpha

-0.73

HUD

-0.72

âĢ¢âĢ¢âĢ¢âĢ¢

-0.72

Lear

-0.70

POSITIVE LOGITS

 assassination

1.38

 assass

1.27

 assassin

1.24

 assassins

1.20

 assassinated

1.18

 Assass

1.13

 assassinate

1.11

 mosqu

0.95

eers

0.87

 murd

0.86

Activations Density 0.015%

mentions of violent actions such as assassination

references to assassination attempts and related discussions

No Comments

No Known Activations