INDEX

Explanations

proper nouns, particularly names of people and places

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

references to specific entities or codes related to health or safety organizations

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

 favour

-0.74

 favor

-0.70

 parting

-0.68

 footing

-0.64

nings

-0.64

 hars

-0.63

 concentrate

-0.62

jud

-0.62

 numbering

-0.62

 retire

-0.62

POSITIVE LOGITS

ECK

1.50

ILD

1.49

APTER

1.49

ANGE

1.48

AMP

1.38

IPS

1.37

INA

1.36

ANG

1.33

AIN

1.30

ANCE

1.28

Activations Density 0.008%

proper nouns, particularly names of people and places

references to specific entities or codes related to health or safety organizations

No Comments

No Known Activations