INDEX

Explanations

function names and programming code patterns

oai_token-act-pair · gpt-3.5-turbo

programming or technical syntax elements

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

advertising

-0.74

lish

-0.61

etheless

-0.61

inately

-0.59

itely

-0.59

rimination

-0.58

MEN

-0.57

Absolutely

-0.56

?????-

-0.55

redibly

-0.55

POSITIVE LOGITS

 Nanto

0.64

 Hubble

0.63

 Malone

0.61

 Toledo

0.61

 Daylight

0.60

 Burke

0.60

 Cutter

0.59

 Shanahan

0.57

 Huntington

0.57

 Kinnikuman

0.57

Activations Density 1.142%

function names and programming code patterns

programming or technical syntax elements

No Comments

No Known Activations