INDEX

Explanations

references to someone or something being alive and well

oai_token-act-pair · gpt-3.5-turbo

phrases indicating the concept of being "alive."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

pled

-0.70

soDeliveryDate

-0.69

ple

-0.66

ciples

-0.66

agall

-0.63

RECT

-0.63

liga

-0.62

iple

-0.62

ples

-0.62

wcsstore

-0.62

POSITIVE LOGITS

lihood

1.11

abouts

0.89

lier

0.86

 alive

0.81

nces

0.74

mares

0.71

beat

0.70

 Alive

0.68

spin

0.66

acebook

0.65

Activations Density 0.032%

references to someone or something being alive and well

phrases indicating the concept of being "alive."

No Comments

No Known Activations