INDEX

Explanations

mentions of government officials or political activities

oai_token-act-pair · gpt-3.5-turbo

mentions of the word "Min" or anything related to the concept of minimalism or measurement

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

enegger

-0.84

fitting

-0.84

atered

-0.81

ä½ľ

-0.73

schild

-0.70

 caliphate

-0.64

 razor

-0.64

hemoth

-0.63

llah

-0.62

fits

-0.62

POSITIVE LOGITS

neapolis

1.22

utes

1.11

ority

0.93

oby

0.92

erva

0.92

otaur

0.91

isters

0.90

ISTER

0.86

nesota

0.84

ister

0.83

Activations Density 0.005%

mentions of government officials or political activities

mentions of the word "Min" or anything related to the concept of minimalism or measurement

No Comments

No Known Activations