INDEX

Explanations

references to people who expose confidential information by publicizing it

oai_token-act-pair · gpt-3.5-turbo

references to whistleblowers and related terms

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

eb

-0.83

pop

-0.77

bal

-0.75

eon

-0.72

road

-0.72

agan

-0.71

alter

-0.70

apo

-0.70

animate

-0.69

anus

-0.69

POSITIVE LOGITS

 whistleblowers

1.20

 whistleblower

1.18

 whistlebl

0.95

 disclosures

0.85

DragonMagazine

0.79

 whistle

0.78

é¾įåĸļå£«

0.78

 revelations

0.74

çīĪ

0.74

ileaks

0.73

Activations Density 0.010%

references to people who expose confidential information by publicizing it

references to whistleblowers and related terms

No Comments

No Known Activations