INDEX

Explanations

references to the word "photo" followed by a number

oai_token-act-pair · gpt-3.5-turbo

references to photography or visual media

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

sheet

-0.76

ruary

-0.70

acious

-0.67

vertisement

-0.66

iments

-0.65

 Lazarus

-0.65

 vacancies

-0.64

innon

-0.64

 resemb

-0.63

 disapp

-0.63

POSITIVE LOGITS

zzi

1.23

veland

0.92

amera

0.88

oga

0.83

ÄŁ

0.82

zzo

0.81

onga

0.81

GP

0.79

ota

0.79

Ã±

0.78

Activations Density 0.015%

references to the word "photo" followed by a number

references to photography or visual media

No Comments

No Known Activations