INDEX

Explanations

phrases related to actions or events that are planned, completed, or managed

oai_token-act-pair · gpt-3.5-turbo

end-of-document markers

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Vaugh

-0.52

ierre

-0.51

gri

-0.49

Kop

-0.49

Eth

-0.49

 Moroc

-0.49

enzie

-0.48

Gap

-0.48

iola

-0.48

 entreprene

-0.46

POSITIVE LOGITS

Â¶

0.69

 hereby

0.65

actionDate

0.65

 Reincarnated

0.64

)?

0.63

0.62

rentices

0.61

)!

0.61

!!!!

0.61

 Released

0.61

Activations Density 0.338%

phrases related to actions or events that are planned, completed, or managed

end-of-document markers

No Comments

No Known Activations