INDEX

Explanations

the word "Peterson"

oai_token-act-pair · gpt-3.5-turbo

mentions of the name "Peterson."

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

 coding

-0.71

code

-0.68

Code

-0.67

 codes

-0.67

codes

-0.67

cho

-0.65

ALTH

-0.64

 slee

-0.63

stri

-0.62

POSITIVE LOGITS

 Peterson

4.03

 Petersen

2.11

Pet

1.68

 Bridgewater

1.46

 Olson

1.39

 Patterson

1.34

 Ramsey

1.28

 Palmer

1.27

 Peters

1.26

 Petra

1.26

Activations Density 0.013%

the word "Peterson"

mentions of the name "Peterson."

No Comments

No Known Activations