INDEX

Explanations

out references to checking something out, such as websites, summaries, or related stories

oai_token-act-pair · gpt-3.5-turbo

phrases that suggest exploring or checking out additional content or resources

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

matter

-0.75

ylum

-0.73

ade

-0.72

amphetamine

-0.71

isma

-0.71

isSpecialOrderable

-0.69

anasia

-0.68

Attempt

-0.65

wig

-0.64

 cumbers

-0.63

POSITIVE LOGITS

our

1.09

 these

0.95

 pics

0.89

 below

0.88

my

0.86

 what

0.85

how

0.84

 whats

0.84

 some

0.84

 this

0.84

Activations Density 0.075%

out references to checking something out, such as websites, summaries, or related stories

phrases that suggest exploring or checking out additional content or resources

No Comments

No Known Activations