INDEX

Explanations

numbers in the hundreds of thousands

oai_token-act-pair · gpt-3.5-turbo

ellipsis or omissions in text

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

Facebook

-0.74

 Turner

-0.63

–

-0.58

Eid

-0.57

Tes

-0.54

–

-0.54

 ÂŃ

-0.53

 Tracker

-0.53

 aforementioned

-0.53

 Greenland

-0.52

POSITIVE LOGITS

....

3.25

.....

2.74

 ....

2.71

......

2.57

.......

2.51

.........

2.42

........

2.22

,...

2.18

 ......

2.13

................

1.96

Activations Density 0.013%

numbers in the hundreds of thousands

ellipsis or omissions in text

No Comments

No Known Activations