INDEX

Explanations

mentions of geographical locations, specifically North Korea and South Korea

oai_token-act-pair · gpt-3.5-turbo

mentions of North Korea and its interactions with South Korea and the United States

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

Interstitial

-0.76

natureconservancy

-0.73

NRS

-0.70

Lens

-0.69

TED

-0.68

 "$:/

-0.64

rant

-0.64

Compan

-0.63

uality

-0.63

itaire

-0.62

POSITIVE LOGITS

 Korea

1.62

 Koreans

1.53

 Korean

1.49

 Jong

1.15

ongyang

1.07

 Vietnamese

1.02

 Pyongyang

0.99

orea

0.94

 Carolina

0.94

 Kurdistan

0.92

Activations Density 0.030%

mentions of geographical locations, specifically North Korea and South Korea

mentions of North Korea and its interactions with South Korea and the United States

No Comments

No Known Activations