INDEX

Explanations

numerical values followed by brief or succinct descriptions

oai_token-act-pair · gpt-3.5-turbo

words associated with high importance or significance in various contexts

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

TPPStreamerBot

-0.30

soDeliveryDate

-0.29

 detrim

-0.29

ecause

-0.29

abase

-0.28

Wan

-0.27

eday

-0.27

SPONSORED

-0.27

 Lumpur

-0.27

 assum

-0.27

POSITIVE LOGITS

Gut

0.30

cool

0.26

 ASUS

0.25

uzz

0.24

htaking

0.24

comed

0.23

ussions

0.23

 political

0.23

 Snowden

0.23

 Cool

0.23

Activations Density 4.226%

numerical values followed by brief or succinct descriptions

words associated with high importance or significance in various contexts

No Comments

No Known Activations