INDEX

Explanations

quotation marks around words or phrases

oai_token-act-pair · gpt-3.5-turbo

conversational turns and responses

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

 nightly

-0.91

 downstream

-0.88

 migr

-0.81

 covenant

-0.78

 agon

-0.77

 clin

-0.76

 funeral

-0.76

asc

-0.75

 stret

-0.75

 disciplines

-0.74

POSITIVE LOGITS

Answer

1.66

Well

1.34

ANS

1.27

Question

1.26

Yes

1.25

Reply

1.23

Absolutely

1.23

JM

1.21

What

1.19

Certainly

1.19

Activations Density 0.233%

quotation marks around words or phrases

conversational turns and responses

No Comments

No Known Activations