INDEX

Explanations

specific numeric patterns or sequences

oai_token-act-pair · gpt-3.5-turbo

numerical data and statistics

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

gets

-0.72

locks

-0.71

BOOK

-0.70

staking

-0.69

¥µ

-0.68

spread

-0.67

 tradem

-0.63

bear

-0.62

 princ

-0.62

mble

-0.59

POSITIVE LOGITS

rd

1.79

RD

1.07

osa

0.88

RF

0.88

rade

0.88

endez

0.87

aquin

0.83

arget

0.82

oint

0.78

Activations Density 0.053%

specific numeric patterns or sequences

numerical data and statistics

No Comments

No Known Activations