INDEX

Explanations

mentions of security, law enforcement, and political parties

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

references to social media content and updates

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 12-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.hook_resid_post

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.11.hook_resid_post

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

 Garfield

-0.73

 Lamar

-0.70

 freshman

-0.70

 Morty

-0.69

 Manson

-0.67

 Oaks

-0.66

 homebrew

-0.66

 Bears

-0.66

ner

-0.63

 Hitchcock

-0.63

POSITIVE LOGITS

Ukraine

1.28

enezuel

1.11

 Telegram

1.03

zbollah

1.02

english

0.99

uterte

0.96

Pakistan

0.96

India

0.95

Syria

0.94

Asia

0.93

Activations Density 0.324%

mentions of security, law enforcement, and political parties

references to social media content and updates

No Comments

No Known Activations