INDEX

Explanations

phrases related to things being hidden or located beneath something

oai_token-act-pair · gpt-3.5-turbo

references to something that is located below a surface or hidden from view

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

fin

-0.64

 strict

-0.61

 Liberties

-0.60

 referrals

-0.59

 Hass

-0.59

adv

-0.58

GST

-0.57

Fin

-0.56

HC

-0.56

VA

-0.56

POSITIVE LOGITS

 beneath

3.21

 underneath

2.75

 below

1.46

below

1.39

neath

1.36

 beside

1.36

 atop

1.33

 Bene

1.30

 behind

1.28

 underside

1.25

Activations Density 0.022%

phrases related to things being hidden or located beneath something

references to something that is located below a surface or hidden from view

No Comments

No Known Activations