INDEX

Explanations

phrases related to law enforcement, government agencies, support services, and warnings

oai_token-act-pair · gpt-3.5-turbo

references to rules, organizations, or regulations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

ividual

-0.82

ancial

-0.77

itably

-0.77

irlf

-0.75

uilt

-0.73

escription

-0.72

;;;;;;;;;;;;

-0.71

othal

-0.70

ogun

-0.70

mercial

-0.67

POSITIVE LOGITS

CLASSIFIED

0.89

 behavi

0.78

 antioxid

0.78

NetMessage

0.77

WAYS

0.76

Balt

0.74

natureconservancy

0.72

GoldMagikarp

0.72

 largeDownload

0.70

 reperc

0.67

Activations Density 3.922%

phrases related to law enforcement, government agencies, support services, and warnings

references to rules, organizations, or regulations

No Comments

No Known Activations