INDEX

Explanations

elements related to newsletter sign-ups

oai_token-act-pair · gpt-3.5-turbo

phrases related to signing up or subscription prompts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ĸļ

-0.82

»Ĵ

-0.78

 nerv

-0.77

amily

-0.71

ãĥīãĥ©ãĤ´ãĥ³

-0.69

ooked

-0.69

Pwr

-0.68

ETHOD

-0.67

 smoot

-0.62

RC

-0.62

POSITIVE LOGITS

ificantly

1.31

atories

1.22

ific

1.18

atory

1.12

 Sign

1.11

atures

1.07

Sign

1.06

ging

1.03

ificant

0.98

ifying

0.95

Activations Density 0.014%

elements related to newsletter sign-ups

phrases related to signing up or subscription prompts

No Comments

No Known Activations