INDEX

Explanations

instances where a user is prompted to input specific information or values

oai_token-act-pair · gpt-3.5-turbo

instructions related to entering data or input fields

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

 Hait

-0.65

cler

-0.64

auld

-0.63

 resil

-0.62

arers

-0.62

unin

-0.61

endez

-0.61

fights

-0.61

Vers

-0.60

stones

-0.60

POSITIVE LOGITS

 username

1.29

 keywords

1.12

 Username

1.11

 desired

1.09

 password

1.09

 credentials

1.06

 captcha

1.05

 name

1.03

URL

1.02

 initials

0.99

Activations Density 0.159%

instances where a user is prompted to input specific information or values

instructions related to entering data or input fields

No Comments

No Known Activations