INDEX

Explanations

phrases related to titles, headlines, captions, and slogans

oai_token-act-pair · gpt-3.5-turbo

components related to article or document titles and headlines

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

ords

-0.60

idis

-0.56

aples

-0.53

existing

-0.53

isters

-0.53

ennes

-0.51

fax

-0.51

leys

-0.51

endo

-0.51

encers

-0.51

POSITIVE LOGITS

"'

1.41

"#

1.39

1.24

"(

1.22

"<

1.20

"$

1.19

"[

1.18

"@

1.17

 "âĢ¦

1.17

 "...

1.16

Activations Density 0.232%

phrases related to titles, headlines, captions, and slogans

components related to article or document titles and headlines

No Comments

No Known Activations