INDEX

Explanations

email confirmation and sign-up related phrases

oai_token-act-pair · gpt-3.5-turbo

phrases related to user sign-up confirmations and email validation

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

forth

-0.66

iann

-0.59

cott

-0.58

 commandments

-0.58

 advoc

-0.57

SPONSORED

-0.57

 humble

-0.56

IDES

-0.55

ides

-0.53

iew

-0.53

POSITIVE LOGITS

 unfocusedRange

0.61

ombat

0.61

BuyableInstoreAndOnline

0.60

 Invalid

0.58

 Login

0.57

 correctly

0.56

leted

0.55

GMT

0.55

orkshire

0.55

Authent

0.54

Activations Density 0.030%

email confirmation and sign-up related phrases

phrases related to user sign-up confirmations and email validation

No Comments

No Known Activations