INDEX

Explanations

letter or symbol sequences that appear to be code or special characters

oai_token-act-pair · gpt-3.5-turbo

special characters or punctuation in textual content

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

ividual

-0.50

wered

-0.38

ulous

-0.37

 lasts

-0.36

 costs

-0.35

halla

-0.35

iland

-0.33

 fulfil

-0.33

 cradle

-0.33

abee

-0.32

POSITIVE LOGITS

NetMessage

0.50

ĵĺ

0.42

PLA

0.39

etheless

0.39

APD

0.39

natureconservancy

0.38

 WATCHED

0.38

terness

0.38

ï¸ı

0.38

Activations Density 2.563%

letter or symbol sequences that appear to be code or special characters

special characters or punctuation in textual content

No Comments

No Known Activations